CN102881006A - Method for splicing and fusing image in multi-projection display system - Google Patents
Method for splicing and fusing image in multi-projection display system Download PDFInfo
- Publication number
- CN102881006A CN102881006A CN2012102747000A CN201210274700A CN102881006A CN 102881006 A CN102881006 A CN 102881006A CN 2012102747000 A CN2012102747000 A CN 2012102747000A CN 201210274700 A CN201210274700 A CN 201210274700A CN 102881006 A CN102881006 A CN 102881006A
- Authority
- CN
- China
- Prior art keywords
- image
- space
- carry out
- projection
- utilize
- 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
Images
Abstract
The invention discloses a method for splicing and fusing images in a multi-projection display system and relates to a method for processing images. The method comprises the following steps: preprocessing an image sequence; geometrically calibrating; spatially splicing images; carrying out preprocessing by using wavelet transform; extracting features; and extracting features of adjacent images and fusing the images. With the adoption of the method, an overlapped part of the images can be eliminated, the sizes of the images are increased, the completeness of frames are guaranteed to a large extent, the revealed image resolution is increased, the projection distance is shortened, and the layering of the frames is increased. The invention provides a registering algorithm which can be used for matching rapidly and stably based on feature points so as to carry out the projection splicing and fusion, so that the method is applicable to various screen types.
Description
Technical field
The present invention relates to a kind of disposal route of image.
Background technology
Fast-developing virtual reality technology in recent years, projection splicing and integration technology have been shifted onto scientific and technological forward position, although but traditional separate unit general projectors low price, but resolution is lower, can not satisfy people's needs, and high-end optical projector is expensive, can not obtain common application, so splicing and integration technology based on many projector are arisen at the historic moment, and have both solved price problem, have also improved resolution.Geometry edge between large screen projection ubiquity projector view field does not mate inconsistence problems at present.This is because giant-screen all is projected pictures by the projection of many projectors usually to be spliced, because in multi-channel projection display system, every projector all is separate, has the slit between the picture that two projectors launch.Although can become very little through adjusting this slit, concerning the professional user, such slit also is unacceptable.And a lot of passage giant-screens are not the plane, but column ring curtain or ball curtain will produce distortion of projection at curved surface like this.So for these giant-screens, not only need to process through professional Fusion Edges, process but also will correct through special surface geometry.
Summary of the invention
The technical problem to be solved in the present invention provides Image Mosaics and the fusion method in a kind of multi-projector display system.
The scheme of technical solution problem of the present invention is by the camera acquisition data, tries to achieve the corresponding relation of wanting the image taken by video camera on projected image and the screen, and these geometric relationships are demarcated in pre-service, carry out geometry correction, obtains the overlapping region.
The geometric calibration method of the BPCA that changes based on small echo comprises following a few step:
(1) adopt histogram equalization that image sequence is carried out pre-service;
(2) geometric calibration by the camera collection data, uses the method for computer vision, demarcates relevant geometrical correspondence; By the method for pure software, get access to the overlapping region seam splicing of many projections;
(3) based on the calibration technique of image space splicing, be mapped to same image space by the image space with a plurality of video cameras;
(4) utilize wavelet transformation to carry out pre-service;
(5) carry out feature extraction with the rectangle split image;
(6) utilize the method for step (5) to extract the feature of adjacent image;
(7) image co-registration.
The present invention can get access to the lap of projector projection, processes by the combination of software and hardware, eliminates lap, increase the size of image, guaranteed to a great extent the integrality of picture, the part of every projector projects entire image, the image resolution ratio that represents like this has been enhanced; Shorten projector distance, increased the stereovision of picture.The invention provides a kind of registration Algorithm based on Feature Points Matching that can be quick, sane and carry out splicing and the fusion of projection, adapt to various screen types, it can be very coarse metope, it also can be special metal screen, not only be applicable to the screen on large plane, also be applicable to the ring curtain, even also be applicable to some irregular a little bent curtains.
Description of drawings
Fig. 1 is the splicing and fusion hardware platform synoptic diagram based on the many projections of BPCA of small echo variation;
Fig. 2 is based on many projections of BPCA splicing of small echo variation and the high-level schematic functional block diagram of integration technology;
Fig. 3 is the overlapping region synoptic diagram that obtains projection;
Fig. 4 is the wavelet decomposition synoptic diagram;
Fig. 5 is that overlapping area image merges synoptic diagram.
Embodiment
The invention will be further described below in conjunction with accompanying drawing.
Fig. 1: the splicing and the hardware platform of fusion of the many projections of BPCA that change based on small echo, the corresponding video camera of PC, simultaneously projector of correspondence, an and projection screen.
Fig. 2: get access to image sequence by video camera, the image sequence that gets access to is carried out pre-service, the present invention carries out histogram equalization as shown in Figure 3 to the original image sequence, meeting finds that the tonal range of image has enlarged or intensity profile is even, thereby increased contrast, make image detail clear, reach the purpose of enhancing.Step 100 pair image carries out wavelet transformation, and along with the increase of wavelet decomposition progression, the dimension of image vector has reduced greatly.Extract low frequency sub-band after the wavelet decomposition and carry out the piecemeal pivot analysis and extract eigenwert and obtain proper vector, the coincidence that utilizes at last nearest neighbor classifier to classify to obtain adjacent image is regional.
The geometric calibration algorithm of the BPCA that changes based on small echo, concrete steps are as follows:
(1) image sequence is carried out pre-service, pre-service generally can be considered illumination, yardstick, the interference of the factors such as change in location.Preprocess method commonly used has nonlinear smoothing filtering, the gray-scale value of pixel is approached, thereby eliminate isolated noise spot; The normalization of image is divided into geometrical normalization and gray scale normalization, and the present invention has adopted histogram equalization, the tonal range that can find image enlarged or intensity profile even, thereby increased contrast, make image detail clear, reach the purpose of enhancing.
(2) geometric calibration.Mainly be by the camera collection data, use the method for computer vision, demarcate relevant geometrical correspondence; By the method for pure software, get access to the overlapping region seam splicing of many projections.Comprise and obtain uncalibrated image space and projector space: PC is linked to each other with projector, and the content that PC is drawn is consistent with the content on throwing into screen, and the content that PC is drawn is kept at buffer area, and this zone is become projector space.Catch projector with camera and throw into content on the screen, the content of projection has a 2D coordinate at the picture of taking, and defining this 2D space is image space.What the 2D coordinate of image space represented is this coordinate on the picture that camera is taken.It is different clearly catching the image space that obtains at diverse location.Sometimes whole system can't be clapped complete a position, therefore need to catch picture from diverse location proofreaies and correct, we need to be corrected to the image space of several diverse locations a unified image space this situation, and we represent with the virtual image space.
Point on the PC screen and the point on the projection screen are corresponding, and point and the point on the PC screen on the image space that the image on the video camera photographed screen obtains also are one to one.Order
m 1Be the point in the projector space
XCoordinate,
m 2For the point
XAfter being thrown into that curtain is upper and being taken by camera,
At the coordinate of image space,
m 1With
m 2Between relation can use function
B 1Be expressed as:
m 2 =
B 1 m 1
Changed by reversible:
m 1=
B 1 -1 m 2
The calibration technique of (3) splicing based on image space: be mapped to same image space by the image space with a plurality of video cameras.Imaging model according to viewpoint obtains:
, wherein
The expression camera is perspective matrixes in the mapping relations of the image space of two positions.By in the feature point extraction Algorithm for Solving following formula
With
(4) utilize wavelet transformation to carry out the reduced complexity that dimension that pre-service wavelet decomposition algorithm reduced image makes subsequent algorithm, simultaneously can information extraction.
Can regard image as binary function
f(x, y), its two-dimensional wavelet transformation is defined as:
In the following formula
jThe expression decomposition scale represents three different high fdrequency components.
If scaling function
And wavelet function
Corresponding filter coefficient matrix is respectively
H(low-pass filter) and
G(Hi-pass filter), original image
Be designated as
C 0, then the 2-d wavelet decomposition algorithm can be described as:
Wherein,
h,
v,
dRepresent respectively level, vertical and diagonal components,
H* and
G* be respectively
HWith
GAssociate matrix.
GBe high frequency (details) composition of image, play the difference effect;
HBe the low frequency composition of image, play smoothing effect.
In wavelet transform, signal decomposition can be become low-frequency information and the high-frequency information of different resolution.Figure below is a wavelet decomposition synoptic diagram, after picture signal process one deck wavelet decomposition, will obtain 4 sub-band images.Low-frequency information is to change slowly part, is the framework of image, also is profile, accounts for the major part of full detail; High-frequency information is to change rapidly part, and it reflects detailed information, accounts for the fraction of full detail.If to sub-band images
LLCarry out wavelet decomposition again, can obtain for the second time four sub-band images again, namely original image has been done wavelet decomposition twice, computational complexity can descend.Depend on the circumstances and also can carry out more frequently wavelet decomposition.What the below showed is that a 2D signal (image) successively is decomposed into approximate component
LL, the level detail component
LH, the vertical detail component
HLWith the diagonal detail component
HH
Select Daub (2) that image is carried out twice decomposition, because wavelet transformation can both provide good local message in spatial domain and frequency domain, through the dimension reduction of the image after the wavelet decomposition, then reduced complexity in computing.Sub-band images after the wavelet decomposition is not so responsive to illumination, angle and trickle expression shape change, has improved the speed of splicing.Therefore, the splicing pretreatment stage at image can decompose image by wavelet transformation.
(5) the present invention uses the rectangle split image to carry out feature extraction, and concrete steps are as follows:
The size of supposing image is
N 
MPixel can be divided into it according to the mode that rectangle is cut apart
rIndividual size is
Sub-block
Then the subimage matrix of all image patterns is regarded as the training sample image vector and implemented the PCA method.Suppose total
Width of cloth image, wherein
iThe width of cloth through the sub-block collection that obtains behind the piecemeal is
,
, all training sample submatrix averages are like this:
Every block of image
With the average image
Difference be:
Then covariance matrix can be expressed as:
Wherein
Utilize svd theorem (SVD) to ask for
C jEigenwert and eigenvectors matrix
, eigenwert is pressed from big to small arrangement, before getting
mIndividual eigenwert characteristic of correspondence vector, here
mBy
Determine, usually
Determine the training submatrix
After the individual orthogonal vector, image that can each is known
Be mapped on the subspace that is consisted of by characteristic image and obtain
mDimensional vector:
Arbitrary width of cloth image wherein
Feature can be trained by each the vector combination of the feature space of submatrix
Expression.According to the contribution situation of the subgraph behind the piecemeal to the entire image discrimination, want degree to give different weights to them
, contribute larger weights larger.Sub-block is carried out the PCA method, obtain eigenwert and proper vector, certainly also need sub-block is projected to the subspace that proper vector is opened, obtain projection coefficient, obtain projection coefficient with multiplied by weight
:
In the following formula
Choose the discrimination of integral body produced larger impact.That a piece of the large representative of weights is important, if give the little weights of part and parcel, can be that whole discrimination reduces.The subimage that the present invention obtains piecemeal all carries out the discrimination that arrives of PCA method identification, judges their significance level according to discrimination.
Weight coefficient is defined as:
Be
The discrimination that the piece sub-block utilizes the PCA method to obtain,
Be block count.Obtain comprehensive projection coefficient by following formula, the recycling piecemeal mates.
(6) utilize method described above to extract the feature of adjacent image, we can obtain a series of projection properties matrix
U j =[
j1,
j2,
Jr], (
j=1,2 ..., L),
rBe the piece number that original image is divided, utilize at last nearest neighbor classifier that these features of extracting are carried out Classification and Identification.
Suppose
jThe projection properties matrix of width of cloth image is
U j , the
xThe projection properties matrix of width of cloth image is
U x Utilize nearest neighbor classifier to carry out Classification and Identification.
Nearest neighbor classifier utilizes the minor increment between unknown sample and the known sample to judge and treats which classification is unknown sample belong to, and it is as follows wherein to define the distance metric criterion:
P(
U ji ,
U xl )=||
U ji -
U xl ||
2
Wherein ||
U Ji -
U Xl ||
2Represent the Euclidean distance between two proper vectors.Obtain min
P(
U Ji ,
U Xl ), then
U Ji The
jOf width of cloth picture
iPiece with
U Xl The
xOf width of cloth picture
lThe piece picture coincides.
(7) image co-registration: after getting access to the overlapping region of adjacent picture, adopt the multi-band fusion algorithm to carry out image co-registration.The specific implementation process is as follows:
1) A and B are made up respectively gaussian pyramid, namely carry out successively down-sampled (2 times) and gaussian filtering.
The pyramid ground floor
GA 0,
GB 0Be A, B itself, to obtaining the second layer behind ground floor down-sampled (2 times), the gaussian filtering
GA 1,
GB 1
2) make up the laplacian pyramid.
LA 1,
LB 1For
GA 1,
GB 1, right
GA 1,
GB 1After rising sampling (2 times), carrying out gaussian filtering.Obtain
LA 0,
LB 0
3) utilize pyramid
LAWith
LBThe structure pyramid
LS LSThe size of every one deck with
LAWith
LBCorrespondent equal.
4) utilize
LSReconstruct overlapping region image.Will
LS 1After rising sampling (2 times), carry out gaussian filtering, then will
LSWith
LS 0Stack obtains net result.
Claims (3)
1. Image Mosaics and the fusion method in the multi-projector display system, its feature comprise that following a few step finishes:
(1) adopt histogram equalization that image sequence is carried out pre-service;
(2) geometric calibration by the camera collection data, uses the method for computer vision, demarcates relevant geometrical correspondence; By the method for pure software, get access to the overlapping region seam splicing of many projections;
(3) based on the calibration technique of image space splicing, be mapped to same image space by the image space with a plurality of video cameras;
(4) utilize wavelet transformation to carry out pre-service;
(5) carry out feature extraction with the rectangle split image;
(6) utilize the method for step (5) to extract the feature of adjacent image;
(7) image co-registration.
2. Image Mosaics and the fusion method in the multi-projector display according to claim 1 system is characterized in that concrete steps are:
(1) adopt histogram equalization that image sequence is carried out pre-service;
(2) geometric calibration, by the camera collection data, demarcate relevant geometrical correspondence, get access to the overlapping region seam splicing of many projections, comprise and obtain uncalibrated image space and projector space: PC is linked to each other with projector, the content that PC is drawn is consistent with the content on throwing into screen, the content that PC is drawn is kept at buffer area, this zone is become projector space, catch projector with camera and throw into content on the screen, the content of projection has a 2D coordinate at the picture of taking, defining this 2D space is image space, the 2D coordinate of image space represents is this coordinate on the picture that camera is taken, and the image space of several diverse locations is corrected to a unified image space, represents with the virtual image space;
Point on the PC screen and the point on the projection screen are corresponding, and point and the point on the PC screen on the image space that the image on the video camera photographed screen obtains also are one to one, order
m 1Be the point in the projector space
XCoordinate,
m 2For the point
XAfter being thrown into that curtain is upper and being taken by camera,
At the coordinate of image space,
m 1With
m 2Between relation can use function
B 1Be expressed as:
m 2 =
B 1 m 1
Changed by reversible:
m 1=
B 1 -1 m 2
(3) based on the calibration technique of image space splicing: be mapped to same image space by the image space with a plurality of video cameras, obtain according to the imaging model of viewpoint:
, wherein
The expression camera is perspective matrixes in the mapping relations of the image space of two positions, by in the feature point extraction Algorithm for Solving following formula
With
(4) utilize wavelet transformation to carry out the reduced complexity that dimension that pre-service wavelet decomposition algorithm reduced image makes subsequent algorithm, simultaneously can information extraction,
Can regard image as binary function
f(x, y), its two-dimensional wavelet transformation is defined as:
In the following formula
jThe expression decomposition scale,
Represent three different high fdrequency components;
If scaling function
And wavelet function
Corresponding filter coefficient matrix is respectively
HWith
G, original image
Be designated as
C 0, then the 2-d wavelet decomposition algorithm can be described as:
Wherein,
h,
v,
dRepresent respectively level, vertical and diagonal components,
H* and
G* be respectively
HWith
GAssociate matrix,
GBe the high frequency composition of image, play the difference effect;
HBe the low frequency composition of image, play smoothing effect;
In wavelet transform, signal decomposition can be become low-frequency information and the high-frequency information of different resolution, after picture signal process one deck wavelet decomposition, to obtain 4 sub-band images, low-frequency information is to change slowly part, is the framework of image, also be profile, account for the major part of full detail; High-frequency information is to change rapidly part, and it reflects detailed information, accounts for the fraction of full detail, if to sub-band images
LLCarry out wavelet decomposition again, can obtain for the second time four sub-band images again, namely original image has been done wavelet decomposition twice, computational complexity can descend, and depends on the circumstances and also can carry out more frequently wavelet decomposition; What the below showed is that a 2D signal successively is decomposed into approximate component
LL, the level detail component
LH, the vertical detail component
HLWith the diagonal detail component
HH
By wavelet transformation image is carried out twice decomposition;
(5) the present invention uses the rectangle split image to carry out feature extraction, and concrete steps are as being to suppose that the size of image is
N 
MPixel can be divided into it according to the mode that rectangle is cut apart
rIndividual size is
Sub-block
, then the subimage matrix of all image patterns is regarded as the training sample image vector and implemented the PCA method, suppose total
Width of cloth image, wherein
iThe width of cloth through the sub-block collection that obtains behind the piecemeal is
,
, all training sample submatrix averages are like this:
Every block of image
With the average image
Difference be:
Then covariance matrix can be expressed as:
Wherein
Utilize the svd theorem to ask for
C jEigenwert and eigenvectors matrix
, eigenwert is pressed from big to small arrangement, before getting
mIndividual eigenwert characteristic of correspondence vector, here
mBy
Determine, usually
,
Determine the training submatrix
After the individual orthogonal vector, image that can each is known
Be mapped on the subspace that is consisted of by characteristic image and obtain
mDimensional vector:
Arbitrary width of cloth image wherein
Feature can be trained by each the vector combination of the feature space of submatrix
Expression; According to the contribution situation of the subgraph behind the piecemeal to the entire image discrimination, want degree to give different weights to them
, contribute larger weights larger, sub-block is carried out the PCA method, obtain eigenwert and proper vector, certainly also need sub-block is projected to the subspace that proper vector is opened, obtain projection coefficient, obtain projection coefficient with multiplied by weight
:
In the following formula
Choose the discrimination of integral body produced larger impact; That a piece of the large representative of weights is important, if give the little weights of part and parcel, can be that whole discrimination reduces, the subimage that piecemeal is obtained all carry out the identification of PCA method to discrimination, judge their significance level according to discrimination;
Weight coefficient is defined as:
Be
The discrimination that the piece sub-block utilizes the PCA method to obtain,
Be block count, obtain comprehensive projection coefficient by following formula, the recycling piecemeal mates;
(6) utilize above-mentioned method to extract the feature of adjacent image, obtain a series of projection properties matrix
U j =[
j1,
j2,
Jr], (
j=1,2 ..., L),
rBe the piece number that original image is divided, utilize at last nearest neighbor classifier that these features of extracting are carried out Classification and Identification;
Suppose
jThe projection properties matrix of width of cloth image is
U j , the
xThe projection properties matrix of width of cloth image is
U x , utilize nearest neighbor classifier to carry out Classification and Identification;
Nearest neighbor classifier utilizes the minor increment between unknown sample and the known sample to judge and treats which classification is unknown sample belong to, and it is as follows wherein to define the distance metric criterion:
P(
U ji ,
U xl )=||
U ji -
U xl ||
2
Wherein ||
U Ji -
U Xl ||
2Represent the Euclidean distance between two proper vectors, obtain min
P(
U Ji ,
U Xl ), then
U Ji The
jOf width of cloth picture
iPiece with
U Xl The
xOf width of cloth picture
lThe piece picture coincides;
(7) image co-registration: after getting access to the overlapping region of adjacent picture, adopt the multi-band fusion algorithm to carry out image co-registration.
3. Image Mosaics and the fusion method in the multi-projector display according to claim 2 system is characterized in that the specific implementation process of image co-registration in this way:
(1) A and B are made up respectively gaussian pyramid, namely carry out successively 2 times of down-sampled and gaussian filterings,
The pyramid ground floor
GA 0,
GB 0Be A, B itself, obtain the second layer behind, the gaussian filtering down-sampled to 2 times of ground floors
GA 1,
GB 1
(2) make up the laplacian pyramid,
LA 1,
LB 1For
GA 1,
GB 1, right
GA 1,
GB 12 times rise sampling after, carrying out gaussian filtering, obtain
LA 0,
LB 0
(3) utilize pyramid
LAWith
LBThe structure pyramid
LS,
LSThe size of every one deck with
LAWith
LBCorrespondent equal;
(4) utilize
LSReconstruct overlapping region image will
LS 12 times rise sampling after, carry out gaussian filtering, then will
LSWith
LS 0Stack obtains net result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012102747000A CN102881006A (en) | 2012-08-03 | 2012-08-03 | Method for splicing and fusing image in multi-projection display system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012102747000A CN102881006A (en) | 2012-08-03 | 2012-08-03 | Method for splicing and fusing image in multi-projection display system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102881006A true CN102881006A (en) | 2013-01-16 |
Family
ID=47482319
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2012102747000A Pending CN102881006A (en) | 2012-08-03 | 2012-08-03 | Method for splicing and fusing image in multi-projection display system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102881006A (en) |
Cited By (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104038714A (en) * | 2013-07-06 | 2014-09-10 | 成都智慧星球科技有限公司 | Projector network intelligent upgrading module and control platform system |
| CN104182951A (en) * | 2014-08-15 | 2014-12-03 | 张建伟 | Multiband image stitching method based on panorama of dual cameras |
| CN104820964A (en) * | 2015-04-17 | 2015-08-05 | 深圳华侨城文化旅游科技有限公司 | Multi-projection-based image splicing and fusing method and system |
| CN105979168A (en) * | 2016-07-18 | 2016-09-28 | 合肥盈川信息技术有限公司 | Novel video splicing data distribution system |
| CN106101581A (en) * | 2016-07-18 | 2016-11-09 | 合肥盈川信息技术有限公司 | A kind of three-dimensional system of coordinate video-splicing data distributing method |
| CN106101580A (en) * | 2016-07-18 | 2016-11-09 | 合肥盈川信息技术有限公司 | A kind of three-dimensional system of coordinate video-splicing multiwindow data distributing method |
| CN106131458A (en) * | 2016-07-18 | 2016-11-16 | 合肥盈川信息技术有限公司 | A kind of new three-dimensional system of coordinate video-splicing multiwindow data distributing method |
| CN106168893A (en) * | 2016-07-18 | 2016-11-30 | 合肥盈川信息技术有限公司 | A kind of video-splicing data distribution system |
| CN106210563A (en) * | 2016-07-18 | 2016-12-07 | 合肥盈川信息技术有限公司 | A kind of three-dimensional system of coordinate video-splicing data distribution system |
| CN106201405A (en) * | 2016-07-18 | 2016-12-07 | 合肥盈川信息技术有限公司 | A kind of novel three-dimensional coordinate system video data distribution system |
| CN106210566A (en) * | 2016-07-18 | 2016-12-07 | 合肥盈川信息技术有限公司 | A kind of video-splicing multiwindow data distributing method |
| CN106447605A (en) * | 2016-10-13 | 2017-02-22 | 中北大学 | Bronze ware three-dimensional fragment virtual splicing method based on PCA (principal component analysis) and FFT (fast Fourier transform) |
| CN106713741A (en) * | 2016-11-16 | 2017-05-24 | 深圳六滴科技有限公司 | Quality diagnosis method and apparatus of panoramic video |
| CN106780616A (en) * | 2016-11-23 | 2017-05-31 | 安徽慧视金瞳科技有限公司 | A kind of projector calibrating method based on the mapping of many matrixes |
| CN107784285A (en) * | 2017-10-26 | 2018-03-09 | 中国电子科技集团公司第五十四研究所 | A kind of the army and the people attribute automatic distinguishing method of optical remote sensing image Ship Target |
| CN109523462A (en) * | 2018-11-14 | 2019-03-26 | 北京奇艺世纪科技有限公司 | A kind of acquisition methods and device of VR video screenshotss image |
| US10297009B2 (en) | 2014-12-22 | 2019-05-21 | Interdigital Ce Patent Holdings | Apparatus and method for generating an extrapolated image using a recursive hierarchical process |
| CN109800765A (en) * | 2017-11-17 | 2019-05-24 | 三橡股份有限公司 | A kind of forming process sampling check for quality method of ocean oil hose |
| CN111260542A (en) * | 2020-01-17 | 2020-06-09 | 中国电子科技集团公司第十四研究所 | SAR image splicing method based on sub-block registration |
| CN111754380A (en) * | 2020-06-28 | 2020-10-09 | 苏州科达科技股份有限公司 | Method and device for positioning image tampered area and storage medium |
| CN112800727A (en) * | 2021-04-14 | 2021-05-14 | 北京三维天地科技股份有限公司 | Method for annotating PDF file and application system |
| CN114294780A (en) * | 2021-11-22 | 2022-04-08 | 丁家智 | Cloud online central air conditioner fault analysis system |
| WO2022121686A1 (en) * | 2020-12-11 | 2022-06-16 | 深圳光峰科技股份有限公司 | Projection fusion method, projection fusion system and computer-readable storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1567376A (en) * | 2003-07-03 | 2005-01-19 | 马堃 | On-site panoramic imagery method of digital imaging device |
| CN1841188A (en) * | 2005-03-30 | 2006-10-04 | 北京大学 | Automatic spherical panorama generation method |
| CN101017315A (en) * | 2007-02-13 | 2007-08-15 | 上海水晶石信息技术有限公司 | Multiscreen playing automatically integrating pretreatment method suitable for irregular screen |
| US20090002394A1 (en) * | 2007-06-29 | 2009-01-01 | Microsoft Corporation | Augmenting images for panoramic display |
-
2012
- 2012-08-03 CN CN2012102747000A patent/CN102881006A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1567376A (en) * | 2003-07-03 | 2005-01-19 | 马堃 | On-site panoramic imagery method of digital imaging device |
| CN1841188A (en) * | 2005-03-30 | 2006-10-04 | 北京大学 | Automatic spherical panorama generation method |
| CN101017315A (en) * | 2007-02-13 | 2007-08-15 | 上海水晶石信息技术有限公司 | Multiscreen playing automatically integrating pretreatment method suitable for irregular screen |
| US20090002394A1 (en) * | 2007-06-29 | 2009-01-01 | Microsoft Corporation | Augmenting images for panoramic display |
Non-Patent Citations (4)
| Title |
|---|
| SONGCAN CHEN等: "Subpattern-Based Principal Component Analysis", 《PATTERN RECOGNITION》 * |
| 丁小丽: "图像拼接技术研究", 《万方学位论文数据库》 * |
| 林克正: "基于2D-DWT与FLD的人脸识别算法", 《计算机工程》 * |
| 陈亮: "基于摄像头校正的多投影仪拼接研究与实现", 《万方学位论文数据库》 * |
Cited By (28)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104038714A (en) * | 2013-07-06 | 2014-09-10 | 成都智慧星球科技有限公司 | Projector network intelligent upgrading module and control platform system |
| CN104182951A (en) * | 2014-08-15 | 2014-12-03 | 张建伟 | Multiband image stitching method based on panorama of dual cameras |
| US10297009B2 (en) | 2014-12-22 | 2019-05-21 | Interdigital Ce Patent Holdings | Apparatus and method for generating an extrapolated image using a recursive hierarchical process |
| CN104820964A (en) * | 2015-04-17 | 2015-08-05 | 深圳华侨城文化旅游科技有限公司 | Multi-projection-based image splicing and fusing method and system |
| CN104820964B (en) * | 2015-04-17 | 2018-03-27 | 深圳华侨城文化旅游科技股份有限公司 | Based on the image mosaic fusion method projected more and system |
| CN106210563A (en) * | 2016-07-18 | 2016-12-07 | 合肥盈川信息技术有限公司 | A kind of three-dimensional system of coordinate video-splicing data distribution system |
| CN106131458A (en) * | 2016-07-18 | 2016-11-16 | 合肥盈川信息技术有限公司 | A kind of new three-dimensional system of coordinate video-splicing multiwindow data distributing method |
| CN106168893A (en) * | 2016-07-18 | 2016-11-30 | 合肥盈川信息技术有限公司 | A kind of video-splicing data distribution system |
| CN106201405A (en) * | 2016-07-18 | 2016-12-07 | 合肥盈川信息技术有限公司 | A kind of novel three-dimensional coordinate system video data distribution system |
| CN106210566A (en) * | 2016-07-18 | 2016-12-07 | 合肥盈川信息技术有限公司 | A kind of video-splicing multiwindow data distributing method |
| CN106101580A (en) * | 2016-07-18 | 2016-11-09 | 合肥盈川信息技术有限公司 | A kind of three-dimensional system of coordinate video-splicing multiwindow data distributing method |
| CN106101581A (en) * | 2016-07-18 | 2016-11-09 | 合肥盈川信息技术有限公司 | A kind of three-dimensional system of coordinate video-splicing data distributing method |
| CN105979168A (en) * | 2016-07-18 | 2016-09-28 | 合肥盈川信息技术有限公司 | Novel video splicing data distribution system |
| CN106447605B (en) * | 2016-10-13 | 2019-05-10 | 中北大学 | The virtual joining method of bronze ware three dimensional fragment based on PCA and FFT |
| CN106447605A (en) * | 2016-10-13 | 2017-02-22 | 中北大学 | Bronze ware three-dimensional fragment virtual splicing method based on PCA (principal component analysis) and FFT (fast Fourier transform) |
| CN106713741A (en) * | 2016-11-16 | 2017-05-24 | 深圳六滴科技有限公司 | Quality diagnosis method and apparatus of panoramic video |
| CN106780616A (en) * | 2016-11-23 | 2017-05-31 | 安徽慧视金瞳科技有限公司 | A kind of projector calibrating method based on the mapping of many matrixes |
| CN107784285A (en) * | 2017-10-26 | 2018-03-09 | 中国电子科技集团公司第五十四研究所 | A kind of the army and the people attribute automatic distinguishing method of optical remote sensing image Ship Target |
| CN107784285B (en) * | 2017-10-26 | 2020-06-09 | 中国电子科技集团公司第五十四研究所 | Method for automatically judging civil and military attributes of optical remote sensing image ship target |
| CN109800765A (en) * | 2017-11-17 | 2019-05-24 | 三橡股份有限公司 | A kind of forming process sampling check for quality method of ocean oil hose |
| CN109523462A (en) * | 2018-11-14 | 2019-03-26 | 北京奇艺世纪科技有限公司 | A kind of acquisition methods and device of VR video screenshotss image |
| CN111260542A (en) * | 2020-01-17 | 2020-06-09 | 中国电子科技集团公司第十四研究所 | SAR image splicing method based on sub-block registration |
| CN111754380A (en) * | 2020-06-28 | 2020-10-09 | 苏州科达科技股份有限公司 | Method and device for positioning image tampered area and storage medium |
| WO2022000861A1 (en) * | 2020-06-28 | 2022-01-06 | 苏州科达科技股份有限公司 | Image tampering area locating method and apparatus, and storage medium |
| WO2022121686A1 (en) * | 2020-12-11 | 2022-06-16 | 深圳光峰科技股份有限公司 | Projection fusion method, projection fusion system and computer-readable storage medium |
| CN112800727A (en) * | 2021-04-14 | 2021-05-14 | 北京三维天地科技股份有限公司 | Method for annotating PDF file and application system |
| CN112800727B (en) * | 2021-04-14 | 2021-07-20 | 北京三维天地科技股份有限公司 | Method for annotating PDF file and application system |
| CN114294780A (en) * | 2021-11-22 | 2022-04-08 | 丁家智 | Cloud online central air conditioner fault analysis system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102881006A (en) | Method for splicing and fusing image in multi-projection display system | |
| Ikeuchi | Computer vision: A reference guide | |
| Winlock et al. | Toward real-time grocery detection for the visually impaired | |
| JP6416208B2 (en) | Near-infrared guide image noise removal | |
| JP5542889B2 (en) | Image processing device | |
| US20070273842A1 (en) | Method And Apparatus For Inhibiting A Subject's Eyes From Being Exposed To Projected Light | |
| WO2020134528A1 (en) | Target detection method and related product | |
| JP6428266B2 (en) | COLOR CORRECTION DEVICE, COLOR CORRECTION METHOD, AND COLOR CORRECTION PROGRAM | |
| CN105765966A (en) | Bowl-shaped imaging system | |
| CN107369159B (en) | Threshold segmentation method based on multi-factor two-dimensional gray level histogram | |
| CN102024155B (en) | Rapid matching method of multispectral images based on edge detection | |
| EP1955282A2 (en) | Adjusting digital image exposure and tone scale | |
| CN110414308B (en) | Target identification method for dynamic foreign matters on power transmission line | |
| CN108154491B (en) | Image reflection eliminating method | |
| Hasinoff | Saturation (imaging) | |
| CN109118544A (en) | Synthetic aperture imaging method based on perspective transform | |
| Bulbul et al. | Human action recognition based on dmms, hogs and contourlet transform | |
| Anantrasirichai et al. | Terrain classification from body-mounted cameras during human locomotion | |
| CN105830091A (en) | Systems and methods for generating composite images of long documents using mobile video data | |
| Mi et al. | Research on a fast human-detection algorithm for unmanned surveillance area in bulk ports | |
| WO2023019793A1 (en) | Determination method, cleaning robot, and computer storage medium | |
| Tan | Specularity, specular reflectance | |
| Fisher | Subpixel estimation | |
| CN115937021A (en) | Polarization defogging method based on frequency domain feature separation and iterative optimization of atmospheric light | |
| Slossberg et al. | Freehand Laser Scanning Using Mobile Phone. |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20130116 |