CN108243333A - The projecting method and back projection method of panoramic video and picture - Google Patents

The projecting method and back projection method of panoramic video and picture Download PDF

Info

Publication number
CN108243333A
CN108243333A CN201611209451.1A CN201611209451A CN108243333A CN 108243333 A CN108243333 A CN 108243333A CN 201611209451 A CN201611209451 A CN 201611209451A CN 108243333 A CN108243333 A CN 108243333A
Authority
CN
China
Prior art keywords
image
spherical surface
block
region
face
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
Application number
CN201611209451.1A
Other languages
Chinese (zh)
Inventor
金怡泽
仲伟昆
其他发明人请求不公开姓名
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shanghai Pai Micro Electronic Technology Co Ltd
Original Assignee
Shanghai Pai Micro Electronic Technology Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shanghai Pai Micro Electronic Technology Co Ltd filed Critical Shanghai Pai Micro Electronic Technology Co Ltd
Priority to CN201611209451.1A priority Critical patent/CN108243333A/en
Publication of CN108243333A publication Critical patent/CN108243333A/en
Pending legal-status Critical Current

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N9/00—Details of colour television systems
    • H04N9/12—Picture reproducers
    • H04N9/31—Projection devices for colour picture display, e.g. using electronic spatial light modulators [ESLM]
    • G—PHYSICS
    • G02—OPTICS
    • G02B—OPTICAL ELEMENTS, SYSTEMS OR APPARATUS
    • G02B13/00—Optical objectives specially designed for the purposes specified below
    • G02B13/06—Panoramic objectives; So-called "sky lenses" including panoramic objectives having reflecting surfaces
    • G—PHYSICS
    • G03—PHOTOGRAPHY; CINEMATOGRAPHY; ANALOGOUS TECHNIQUES USING WAVES OTHER THAN OPTICAL WAVES; ELECTROGRAPHY; HOLOGRAPHY
    • G03B—APPARATUS OR ARRANGEMENTS FOR TAKING PHOTOGRAPHS OR FOR PROJECTING OR VIEWING THEM; APPARATUS OR ARRANGEMENTS EMPLOYING ANALOGOUS TECHNIQUES USING WAVES OTHER THAN OPTICAL WAVES; ACCESSORIES THEREFOR
    • G03B37/00—Panoramic or wide-screen photography; Photographing extended surfaces, e.g. for surveying; Photographing internal surfaces, e.g. of pipe
    • G03B37/04—Panoramic or wide-screen photography; Photographing extended surfaces, e.g. for surveying; Photographing internal surfaces, e.g. of pipe with cameras or projectors providing touching or overlapping fields of view
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
    • G06F3/011—Arrangements for interaction with the human body, e.g. for user immersion in virtual reality
    • G06F3/013—Eye tracking input arrangements
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N9/00—Details of colour television systems
    • H04N9/12—Picture reproducers
    • H04N9/31—Projection devices for colour picture display, e.g. using electronic spatial light modulators [ESLM]
    • H04N9/3141—Constructional details thereof
    • H04N9/3147—Multi-projection systems

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • General Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Human Computer Interaction (AREA)
  • Optics & Photonics (AREA)
  • Image Processing (AREA)

Abstract

The projecting method of projection and back projection method the invention discloses a kind of panoramic video and picture, the panoramic video and picture includes step：By polyhedron to spherical projection, spherical surface image is divided into multiple surfaces；By mapping transformation, the mapping relations between each surface of spherical surface image and polyhedral each face are established respectively；Respectively by the pixel projection in each surface of spherical surface image to there are polyhedral faces of mapping relations.The back projection method of the panoramic video and picture includes step：By polyhedron image to spherical projection, spherical surface is divided into multiple surfaces；By mapping transformation, the mapping relations between each face of polyhedron image and each surface of spherical surface are established respectively；Respectively by the pixel projection in each face of polyhedron image to there are the surfaces of the spherical surface of mapping relations.The present invention is by way of to panoramic picture segmentation and polyhedric projection so that pixel distribution is uniform, caused by inhibition panoramic picture projection the problem of the decline of picture quality.

Description

The projecting method and back projection method of panoramic video and picture
Technical field
The present invention relates to the projecting method and back projection method of a kind of video and picture more particularly to a kind of panoramic video and The projecting method and back projection method of picture.
Background technology
Virtual reality (Virtual Reality, abbreviation VR) technology is presently most popular technical field, particularly heavy Immersion virtual reality (Immersive Virtual Reality), user carries out man-machine friendship by wearing Helmet Mounted Display Mutually, the head position information of user can be fed back to host by the helmet in real time, and host then sends regarding for user perspective direction To display, user arbitrarily can change head position and obtain the scenes of different visual angles frequency, just like it is on the spot in person, reach high The feeling of immersion of degree.
Video, picture browsing under reality environment are different from the two-dimensional flat screens such as traditional television set, mobile phone Browsing, is in due to the use of person under virtual three-dimensional environment, and user can arbitrarily change visual angle and appreciate panorama, virtual In reality system, video or picture browsing are panoramas, and people are stood in the position of the centre of sphere, are attached on entire spherical surface to appreciate Video or image.Therefore, how more accurately by image projection to spherical surface and how by entire spherical panorama picture preserve It is the multimedia system critical issue to be solved under reality environment on to two dimensional surface.
Current panoramic video, the projecting method of picture are mainly derived from traditional panoramic pictures projecting method, mainly adopt The projection of sphere-plane is established with cylindrical surface projection pattern, in addition to this, the seldom panoramic projection mode of another use scope It is cubic projection.Either cylindrical surface projection or cubic projection, are all directly by figure all in entire panoramic picture Picture pixel is directly projected to spherical surface, as shown in Figure 1, when directly projecting all image pixels to spherical surface, the picture on spherical surface Element distribution is there are density, the problem of all leading to the uneven distribution of spherical surface pixel.
Left side is cylindrical surface projection in Fig. 1, it can be seen that nearby pixel distribution is sparse under the line, and near south poles, Pixel distribution is very close, in virtual reality system, if projected using cylindrical surface, then user is under the angle of head-up, it is seen that Image as unit area pixel it is on the low side, and in the case where looking up and overlooking, it is seen that image as unit area pixel it is on the high side.With circle Cylindrical surface projecting is similar, and right side is the spherical surface pixel distribution situation under cubic projection in Fig. 1, it is seen that close to a cube honorable center Position pixel distribution it is sparse, close to the position of cube apex, pixel distribution is intensive.
Image pixel uneven distribution in Fig. 1 can lead to the problem of as follows：
1st, panoramic picture quality declines
In the region that pixel distribution is sparse, picture quality is decreased obviously；In the region that pixel distribution is intensive, due to pixel mistake In intensive, pixel can be caused to be overlapped when texture renders, also result in image quality decrease.
2nd, data redundancy
Uneven due to pixel distribution, image, video subjective vision effect are often by the sparse region institute of pixel distribution It determines.In order to promote visual quality, need to meet sparse region lowest resolution requirement, be resulted in this way in other image-regions Data redundancy.Data redundancy causes the data volume of image/video to become larger, and is caused greatly on data storage, data network transmission Waste.
The often image data after being projected to panoramic picture in the case of above-mentioned image pixel uneven distribution at present It is handled and is unevenly distributed to eliminate image pixel.Such as the ball to be formed is projected by cylindrical surface projection pattern in panoramic picture After the image of face, pixel distribution in spherical surface image is adjusted to spherical surface image procossing.But this processing mode needs increase greatly Image real time transfer is measured, the reduction of encoding-decoding efficiency can be all led in Video coding compression or video decoding, so uncomfortable For HD video or picture.
Invention content
The technical problem to be solved by the present invention is to only be to entirely scheming to overcome the projection of panoramic picture in the prior art As pixel projection caused by pixel distribution is uneven and pixel data redundancy, and then influences the picture quality of panoramic picture The defects of, the projecting method and back projection method of a kind of panoramic video and picture are provided, by panoramic picture segmentation and more The mode of face body projection so that pixel distribution is uniform, caused by inhibition panoramic picture projection the problem of the decline of picture quality.
The present invention is to solve above-mentioned technical problem by following technical proposals：
The present invention provides a kind of panoramic video and the projecting method of picture, its main feature is that, the panoramic video and picture Projecting method include the following steps：
By polyhedron to spherical projection, spherical surface image is divided into multiple surfaces；
By mapping transformation, the mapping relations between each surface of spherical surface image and polyhedral each face are established respectively；
Respectively by the pixel projection in each surface of spherical surface image to there are polyhedral faces of mapping relations.
The projecting method of the present invention is used to implement the image that the content transformation of panoramic video or picture is polyhedral type, Consequently facilitating the coding compression of subsequent image.Wherein spherical diagram seems the packets such as each frame picture or the panoramic pictures of panoramic video Carrier containing panorama image information.
And the pixel projection of the spherical surface image after division is also realized pixel by the present invention to polyhedral each face It is uniformly distributed.Wherein pixel projection can realize being uniformly distributed for pixel using existing pixel projection mode in the present invention.
Preferably, to spherical projection, included during spherical surface image is divided into multiple surfaces by polyhedron with Lower step：
Where polyhedron is set to spherical surface image in sphere, and the point centered on the centre of sphere；
Polyhedral each vertex and seamed edge are mapped into spherical surface image；
Spherical surface image is divided into multiple surfaces by each vertex and seamed edge for mapping to spherical surface image.
Preferably, by polyhedron, to spherical projection, the Cheng Qian that crosses that spherical surface image is divided into multiple surfaces further includes Following steps：
Polyhedron is established, wherein polyhedral surface area is identical with the surface area of sphere where spherical surface image.
During wherein polyhedral each face is mapped on spherical surface in the present invention, pass through polyhedral vertex and rib While the region corresponding to each face of polyhedron is marked off on spherical surface.
And preferable picture can be reached when polyhedral surface area is identical with the surface area of sphere where spherical surface image Element is uniformly distributed.
Preferably, the pixel projection in each surface by spherical surface image is to there are polyhedral faces of mapping relations Include the following steps in the process：
To every a pair, there are the surfaces of the spherical surface image of mapping relations and polyhedral face execution following steps：
By the surface segmentation of spherical surface image into multiple images block；
By mapping transformation, establish between each image block on the surface of spherical surface image deposit on polyhedral face respectively In the block of mapping relations；
It is built in the surface of all spherical surface images there are mapping relations and polyhedral face and completes image block and block After mapping relations, surface and polyhedral face to all spherical surface images there are mapping relations perform following steps：
It will be in the pixel projection to polyhedral face there are mapping relations of image block each in the surface of spherical surface image Block.
Spherical surface image is being divided into multiple regions by the present invention, and the image of each region is mapped to polyhedral each During face, further each region is further divided and mapped, polyhedral image will be mapped to and further refined, more Good realization pixel is uniformly distributed.
Preferably, there are mapping relations between foundation and each image block on the surface of spherical surface image on polyhedral face Block during include the following steps：
Each vertex of image block each in the surface of spherical surface image is mapped into polyhedral face；
By mapping transformation, polyhedral face is divided into multiple pieces by each vertex for mapping to polyhedral face.
Preferably, there are mapping relations between foundation and each image block on the surface of spherical surface image on polyhedral face Block process after iteration perform following steps：
Image block is divided into multiple images region；
By mapping transformation, established on polyhedral piece respectively with each image-region of the image block of spherical surface image it Between there are mapping relations region；
It is built in the surface of all spherical surface images there are mapping relations and polyhedral face and completes image-region and area After the mapping relations in domain, surface and polyhedral face to all spherical surface images there are mapping relations perform following steps：
By the pixel projection of image-region each in the image block of spherical surface image to there are polyhedral pieces of mapping relations In region.
The present invention preferably realizes being uniformly distributed for pixel further by being split to image, especially by repeatedly In generation, repeatedly divides image, i.e., the image-region after constantly mapping segmentation carries out further segmentation and maps, so that It is minimized, such as divide the image into single pixel eventually for the image-region of projection.
It should be noted that the number of iteration can freely be set according to coding and decoding video and the needs of picture quality, And after the completion of iteration, the pixel projection of all image-regions after panoramic picture is cut is to there are the regions of mapping relations In, so as to fulfill panoramic picture to polyhedral projection.
Preferably, it is established on polyhedral piece and there is mapping between each image-region of the image block of spherical surface image Include the following steps during the region of relationship：
Each vertex of each image-region of the image block of spherical surface image is mapped to polyhedral piece；
By mapping transformation, map to polyhedral piece of each vertex and be divided into multiple regions by polyhedral piece.
Preferably, the pixel projection of image block each in the surface of spherical surface image or image-region is being closed to there is mapping Include the following steps during block or region in polyhedral face of system：
According to the top in block or region in the vertex of image block or image-region in the surface of spherical surface image and polyhedral face The mapping relations of point, are projected image pixel in image block or image-region to block or region by pixel value difference algorithm；Or
Calculate in the surface of spherical surface image in each image pixel of image block or image-region and polyhedral face block or The mapping relations of region relative position project image pixel in image block or image-region to block or region.
In the present invention spherical surface is realized using the pixel value difference algorithm in conventional images algorithm and other pixel projection methods Image is to polyhedral projection.
Preferably, during the surface of spherical surface image or image block are divided into multiple images block or image-region, The surface of spherical surface image or image block are evenly divided into multiple images block or image-region.
Preferably, established on polyhedral face with each image block on the surface of spherical surface image or each image-region it Between there are the block of mapping relations or region during, each image block or image-region mapping block or region area phase Deng.
In the present invention even partition image-region or image block in addition to be based on the even partition of area other than, can also be The non-uniform manner of even partition mode or arbitrary mess shape based on angle, in addition non-uniformly distributed load mode be applied equally to The region of the present invention or fast cutting, such as the non-homogeneous side of the non-uniformly distributed load based on quadrilateral mesh or arbitrary mess shape Formula.
Preferably, each piece or region are quadrangle.
Preferably, each piece or region are square.Square is relatively simple convenient projection model, is conducive to improve Calculating speed is projected, and square-shaped image is suitable for existing coding techniques, and square can be in VR visual effects Consistency on keeping lateral and vertical.
Preferably, the mapping transformation converts (Shi Naide transformation), RD transformation, Cube transformation, Unicube changes for Snyder Change, IsoCube transformation, perspective projection transformation or Lambert authalic projections transformation etc..Above-mentioned transform method is that existing mapping becomes Common mapping changing method in changing.
Preferably, the polyhedron is regular polygon.
Preferably, the polyhedron is cube or regular dodecahedron.Cube facilitates expansion, is suitable for existing coding Technology, and cube model is relatively simple projection model, is conducive to improve projection calculating speed, and cube model Provide the pixel distribution being more uniformly distributed.
Preferably, each surface includes multiple images pixel.
Preferably, each image block includes at least one image pixel or each image-region including at least one Image pixel.
That is the limit of present invention segmentation image is that only have single pixel inside each region of segmentation.It should be noted that Be the present invention using image divide by the way of come projected image, so even if being divided into pixel there are one each regions, Core is still that segmentation and projects image, in the prior art between the mode that all pixels of whole image integrally project is cut It is so different.
The present invention also provides a kind of panoramic video and the back projection method of picture, its main feature is that, the panoramic video and The back projection method of picture includes the following steps：
By polyhedron image to spherical projection, spherical surface is divided into multiple surfaces；
By mapping transformation, the mapping relations between each face of polyhedron image and each surface of spherical surface are established respectively；
Respectively by the pixel projection in each face of polyhedron image to there are the surfaces of the spherical surface of mapping relations.
Preferably, it is described by polyhedron image to spherical projection, include during spherical surface is divided into multiple surfaces Following steps：
Where polyhedron image is set to spherical surface in ball, and the point centered on the centre of sphere；
Each vertex polyhedral in polyhedron image and seamed edge are mapped into spherical surface；
Spherical surface is divided into multiple surfaces by each vertex and seamed edge for mapping to spherical surface.
Preferably, by polyhedron to spherical projection, by spherical surface be divided into multiple surfaces cross Cheng Qian further include it is following Step：
Sphere is established, wherein polyhedral surface area is identical with the surface area of sphere spherical surface where polyhedron image.
Preferably, by the pixel projection in each face of polyhedron image to there are the mistakes on the surface of the spherical surface of mapping relations Include the following steps in journey：
To every a pair, there are the faces of the polyhedron image of mapping relations and the surface execution following steps of spherical surface：
The face of polyhedron image is evenly divided into multiple images block；
By mapping transformation, deposited between each image block in the face of foundation and polyhedron image on the surface of spherical surface respectively In the block of mapping relations；
It is built on the face of all polyhedron images there are mapping relations and the surface of spherical surface and completes image block and block After mapping relations, the surface in face and spherical surface to all polyhedron images there are mapping relations performs following steps：
By the pixel projection of image block each in the face of polyhedron image in there are the surface of the spherical surface of mapping relations Block.
Spherical surface is being divided into multiple regions, and polyhedral each face is mapped to the process of each region by the present invention In, further each region is further divided and mapped, the image for mapping to spherical surface is further refined, preferably realizes picture Element is uniformly distributed.
Preferably, the block on the surface of spherical surface between foundation and each image block of polyhedron image there are mapping relations During include the following steps：
Each vertex of image block each on the face of polyhedron image is mapped to the surface of spherical surface；
By mapping transformation, the surface of spherical surface is divided into multiple pieces by each vertex for mapping to the surface of spherical surface.
Preferably, there are mapping relations between foundation and each image block in the face of polyhedron image on the surface of spherical surface Block process after iteration perform following steps：
Image block is divided into multiple images region；
By mapping transformation, established on the block of spherical surface respectively with each image-region of the image block of polyhedron image it Between there are mapping relations region；
It is built on the face of all polyhedron images there are mapping relations and the surface of spherical surface and completes image-region and area After the mapping relations in domain, the face on surface and polyhedron image to all spherical surfaces there are mapping relations performs following steps：
By the pixel projection of image-region each in the image block of polyhedron image to there are the blocks of the spherical surface of mapping relations In region.
The present invention preferably realizes being uniformly distributed for pixel further by being split to image, especially by repeatedly In generation, repeatedly divides image, i.e., constantly the image-region after segmentation is further divided, so that eventually for The image-region of projection minimizes.
Preferably, it is established on the block of spherical surface and there is mapping between each image-region of the image block of polyhedron image Include the following steps during the region of relationship：
Each vertex of each image-region of the image block of polyhedron image is mapped to the block in the surface of spherical surface；
By mapping transformation, the block of spherical surface is divided into multiple areas by each vertex of the block mapped in the surface of spherical surface Domain.
Preferably, the pixel projection of image block each in the face of polyhedron image or image-region is being closed to there is mapping Include the following steps during block or region in the surface of the spherical surface of system：
According to the top in block or region in the surface on the vertex and spherical surface of image block or image-region in the face of polyhedron image The mapping relations of point, are projected image pixel in image block or image-region to block or region by pixel value difference algorithm；Or
Calculate polyhedron image face in image block or image-region each image pixel and spherical surface surface in block or The mapping relations of region relative position project image pixel in image block or image-region to block or region.
Preferably, during the face of polyhedron image or image block are divided into multiple images block or image-region, The face of polyhedron image or image block are evenly divided into multiple images block or image-region.
Preferably, established on the surface of spherical surface with each image block in the face of polyhedron image or each image-region it Between there are the block of mapping relations or region during, the area equation of each image block or image-region.
In the present invention even partition image-region or image block in addition to be based on the even partition of area other than, can also be The non-uniform manner of even partition mode or arbitrary mess shape based on angle, in addition non-uniformly distributed load mode be applied equally to The region of the present invention or fast cutting, such as the non-homogeneous side of the non-uniformly distributed load based on quadrilateral mesh or arbitrary mess shape Formula.
Preferably, each image block or image-region are quadrangle.
Preferably, each image block or image-region are square.Square is relatively simple convenient projection model, is had Calculating speed is projected, and square-shaped image is suitable for existing coding techniques, and square can be regarded in VR conducive to improving Feel the consistency on keeping lateral and vertical in effect.
Preferably, the mapping transformation is Snyder transformation, RD transformation, Cube transformation, Unicube is converted, IsoCube becomes Change, perspective projection transformation or Lambert authalic projections transformation etc..
Preferably, the polyhedron is regular polygon.
Preferably, the polyhedron is cube or regular dodecahedron.Cube facilitates expansion, is suitable for existing coding Technology, and cube model is relatively simple projection model, is conducive to improve projection calculating speed, and cube model Provide more uniform pixel distribution.
Preferably, polyhedral each bread pixel containing multiple images.
Preferably, each image block includes at least one image pixel or each image-region including at least one Image pixel.
On the basis of common knowledge of the art, above-mentioned each optimum condition can be combined arbitrarily to get each preferable reality of the present invention Example.
The positive effect of the present invention is：
The present invention is cut by image and polyhedric projection, realizes that even partition and pixel in spherical surface are uniformly distributed, and And by being applied to projection and back projection, the uniform pixel for not only having solved the problems, such as spherical panorama scene uses, but also solve panorama Uniform pixel distribution problem under playing.So it is played available for pan-shot and panorama, and pixel is held on spherical surface The uniformity of distribution.
Polyhedric projection of the present invention is not only less using face, and perspective plane and inside division are all quadrangle especially squares Shape or square are more suitable for video and image and encode, final compression efficiency also higher.
And the present invention, also by the multiple cutting to figure, to further improve, uniform pixel uses and uniform pixel is divided Cloth.
Description of the drawings
Fig. 1 is the picture element density distribution map that the panoramic pictures of the prior art project on spherical surface.
Fig. 2 is the mapping schematic diagram that the Snyder of the embodiment of the present invention 1 is converted.
Fig. 3 is the schematic diagram of the segmentation projection of the embodiment of the present invention 2.
Fig. 4 is the embodiment of the present invention 2 and drop shadow effect's comparison diagram of the prior art.
Fig. 5 is the schematic diagram of the segmentation projection of the embodiment of the present invention 7.
Fig. 6 is the schematic diagram of the segmentation projection of another embodiment of the present invention.
Fig. 7 is converted to the transformation of rectangle diagram picture for parallelogram image after the image projection of the embodiment of the present invention 7 Schematic diagram.
Specific embodiment
It is further illustrated the present invention below by the mode of embodiment, but does not therefore limit the present invention to the reality It applies among a range.
It is the pixel projection to whole image in panoramic picture projection in the prior art, regardless of cylinder projection is also Be cubic projection can caused by pixel distribution is uneven and pixel data redundancy, and then influences the image matter of panoramic picture Amount.And in order to which pixel distribution caused by inhibiting this projection is uneven and data redundancy and improves picture quality, the prior art Only processing is optimized by the image after being projected to panoramic picture so that pixel is evenly distributed on final spherical surface.And And this subsequent processing can cause largely to calculate, and be unfavorable for the processing of High-speed video images and the place of high clear video image Reason.
The present invention is cut by image and polyhedric projection, is realized during panoramic pictures project in the uniform of spherical surface Segmentation and pixel are uniformly distributed, and not only increase the picture quality after projection, also reduce the calculation amount of subsequent image processing, are fitted For the processing of High-speed video images and the processing of high clear video image.But also by the multiple cutting to figure, into one The uniform pixel that step improves projection stage uses and uniform pixel distribution.
And the image cutting of the present invention can not only be applied to projection with polyhedric projection, be also applied for back projection, both solved The uniform pixel for spherical panorama scene of having determined uses problem, and solves the problems, such as the uniform pixel distribution under panorama broadcasting.So Can be used for spherical uniform pixel sampling, spherical uniform pixel is shown and spherical uniform mesh generation, number may finally be eliminated According to redundancy, to obtain higher Y-PSNR (PSNR) under identical sampling number.
In addition the present invention is not only less using face using polyhedric projection, and perspective plane and inside division are all quadrangles Especially rectangular or square is more suitable for video and image and encodes, final compression efficiency also higher, so as to preferably fit For coding and decoding video.
The present invention is further explained in detail below by following embodiments.
Embodiment 1
The present embodiment is as follows by the flow of the image of spherical surface or the method for projecting video image：
Step S101, cube is established, wherein cubical surface area is identical with the surface area of sphere where spherical surface image. By cube spherical projection, spherical surface image is divided into multiple surfaces.The present embodiment uses cube, arbitrary just more in addition Bian Ti, gengon etc. may be used to the present embodiment.But for calculation amount and effectiveness, positive gengon has including cube Best effectiveness.
Specific segmentation spherical surface image further includes step S1011, is set to polyhedron in the sphere of spherical surface image place, and The point centered on the centre of sphere.
Step S1012, polyhedral each vertex and seamed edge are mapped into spherical surface image.
Spherical surface image is divided into multiple surfaces by each vertex and seamed edge for step S1013, mapping to spherical surface image.
Step S102, it is converted, is established respectively between each surface of spherical surface image and polyhedral each face by Snyder Mapping relations.
For example, obtain mapping relations via Schneider transformation calculations as shown in Fig. 2, reference numeral 301 it is signified [A, B, C, D] it is face on cubic projection to spherical surface, reference numeral 303 signified [A', B', C', D'] is that cube is projected to ball Face on face.
Specific vertex mapping step is as follows：
The radius of spherical surface is wherein set as R, the cube model inscribe radius of a ball is r.By each face of cube according to right Title property is divided into triangular element, and such as the triangular element pointed by reference numeral 304, then corresponding surface is also divided on spherical surface Spherical triangle unit, as indicated by reference numeral 302.
It is as follows based on cubical division：Each face on cube is divided into 8 triangles, then pair on spherical surface Face is answered to be also divided into 8 spherical triangles.As shown in reference numeral 301, by AF, BE, KM and NL four edges, by face 301 It is divided into 8 spherical triangles；As shown by reference number 303, by A'F', B'E', K'M' and N'L' four edges, by face 303 It is divided into 8 triangles.
As indicated by reference numeral 302, for point P on spherical surface, the division region according to where it builds the corresponding pole of point Coordinate (α, t), wherein, α is spherical angle, i.e. the angle of face POI knead doughs AOI；T be spherical distance corresponding to radian, i.e. point P with Radian corresponding to the spherical distance of point I.
As shown in reference numeral 304, for point P' on cube, the division region according to where it can be built corresponding Polar coordinates (θ, ρ), wherein, θ is angle, i.e. angle ∠ A'I'P'；ρ is distance, i.e. the length of line segment P'I'.
Mapping (P → P') of the umbilical point to cube point is established according to following equation, i.e., is mapped spherical polar coordinates (α, t) To cube polar coordinates (θ, ρ), specific formula for calculation is as follows：
H=arccos (sin (α) sin (G) cos (g)-cos (α) cos (G)) (1)
Wherein, r is the inscribe radius of a ball in cube Current projection face,It is the interior angle I ' A ' N ' of plane triangle A ' N ' I ' Radian, G is the radian of the interior angle IAN of spherical triangle ANI, and g is side AI (i.e. orthodrome between point A and point I) corresponding Radian；
Mapping (P' → P) of the polyhedron point to umbilical point is established according to following equation, i.e., is reflected cube polar coordinates (θ, ρ) Spherical polar coordinates (α, t) are mapped to, specific formula for calculation is as follows：
(using formula (1), (2), (3)) is obtained using Newton iteration method in the transformation of polar angle, and wherein S is obtained by formula (3)：
δ α=- F (α)/F'(α) (10)
The initial value of α is set first as α0=θ carries out first time iteration, is calculated
δ α=- F (α0)/F'(α0) (11)
Then the value of α is updated：
α1=α0+δα (12)
An iteration terminates at this time, and the approximate solution of more accurate α is can obtain after several wheel iteration；
And t can be calculated by formula (6)：
T=2arcsin (ρ sin (q/2)/d) (13)
Wherein, variable q and d is obtained according to formula (4) and (5).
Mapping relations between cube and spherical surface can be calculated by above-mentioned formula.
Step S103, respectively by the pixel projection in each surface of spherical surface image to there are the cubical of mapping relations Face.
Multiple pixels are included in the present embodiment in each region of divided spherical surface image, divides in the present embodiment and reflects The uneven of pixel distribution can effectively be inhibited by penetrating the mode of projection.
Embodiment 2
The image of the spherical surface of the present embodiment or the method for projecting video image on the basis of the projecting method of embodiment 1, To spherical projection to cubical face further segmentation and mapping projections, and then further improve the distributing homogeneity of pixel. Specific method flow is as follows：
Step S201, cube is established, wherein cubical surface area is identical with the surface area of sphere where spherical surface image. By cube spherical projection, spherical surface image is divided into multiple surfaces.
Specific segmentation spherical surface image further includes step S2011, is set to polyhedron in the sphere of spherical surface image place, and The point centered on the centre of sphere.
Step S2012, polyhedral each vertex and seamed edge are mapped into spherical surface image.
Spherical surface image is divided into multiple surfaces by each vertex and seamed edge for step S2013, mapping to spherical surface image.
Step S202, it is converted, is established respectively between each surface of spherical surface image and polyhedral each face by Snyder Mapping relations.Its mapping method is identical with step S102 in embodiment 1, and details are not described herein again.
Step S203, every a pair there are the surface of the spherical surface image of mapping relations and polyhedral face is performed and walked as follows Suddenly：
S2031, by the surface segmentation of spherical surface image into multiple images block.
S2032, it is converted by Snyder, establishes each image with the surface of spherical surface image on polyhedral face respectively There are the blocks of mapping relations between block.
Specific segmentation step is as follows：
Each vertex of image block each in the surface of spherical surface image is mapped into polyhedral face.
It is converted by Snyder, polyhedral face is divided into multiple pieces by each vertex for mapping to polyhedral face.
For example, each sublist face to spherical surface 205 is carried out according to each face of cube after suitably dividing.To each of spherical surface Spherical surface is marked off several cut zone by a sub- surface 206 by way of stroke grid lines of such as reference numeral 207.Attached drawing mark S' shown in note 208 is the mesh segmentation region marked off, and what [A', B', C', D'] was characterized is four tops of grid S' Point.
It is converted by Snyder, establishes spherical surface and divide vertex to the mapping relations on polyhedron segmentation vertex.I.e. by attached drawing mark All grid vertexes are all converted by Snyder on Algorithm mapping to the face 202 of cube 201 in sublist face shown in note 206. By taking grid S' as an example, need [A', B', C', D'] four points becoming scaling method on the face of cube 201 202 by Snyder Calculate corresponding [A, B, C, D] coordinate position, i.e., grid S shown in reference numeral 204, that is, calculate [A', B', C', D'] → [A, B, C, D].
According to the mapping relations, spherical surface to cubic projection is exactly to throw the segmentation grid lines of spherical surface shown in reference numeral 207 Cube shown in shadow to reference numeral 203 divides grid lines, and then establishes reflecting for above-mentioned [A', B', C', D'] → [A, B, C, D] Relationship is penetrated, that is, establishes the coordinate mapping relations of S' → S.
It may be noted that the surface of spherical surface image is evenly divided into the identical multiple images block of area in the present embodiment.
Step S204, completion image is built on the surface of all spherical surface images there are mapping relations and polyhedral face After the mapping relations of block and block, surface and polyhedral face to all spherical surface images there are mapping relations perform following step Suddenly：
It will be in the pixel projection to polyhedral face there are mapping relations of image block each in the surface of spherical surface image Block.So spherical surface image, to cubical projection, being exactly will be in the pixel projection on spherical surface in S' to cubical S faces.
Specifically, according to block in the vertex of image block in the surface of spherical surface image and cubical face in the present embodiment The mapping relations on vertex are projected image pixel in image block or image-region to block or region by pixel value difference algorithm.
Embodiment 3
The image of the spherical surface of the present embodiment or the method for projecting video image on the basis of the projecting method of embodiment 2, Divide simultaneously mapping projections again to spherical projection to cubical face, and then improve the distributing homogeneity of pixel.Specific method Flow is as follows：
Step S301, polyhedral each vertex and seamed edge are mapped into spherical surface image.
Spherical surface image is divided into multiple surfaces by each vertex and seamed edge for step S302, mapping to spherical surface image.
Step S303, it is converted, is established respectively between each surface of spherical surface image and polyhedral each face by Snyder Mapping relations.Its mapping method is identical with step S202 in embodiment 2, and details are not described herein again.
Step S304, every a pair there are the surface of the spherical surface image of mapping relations and polyhedral face is performed and walked as follows Suddenly：
S3041, by the surface segmentation of spherical surface image into multiple images block
S3042, it is converted by Snyder, establishes each image with the surface of spherical surface image on polyhedral face respectively There are the blocks of mapping relations between block.
Step S305, mapping is split to each image block again, specifically comprised the following steps：
S3051, by image block into multiple images region.
S3052, it is converted by Snyder, establishes each figure with the image block of spherical surface image on cubical piece respectively As the region there are mapping relations between region.
Specifically, exactly each vertex of each image-region of the image block of spherical surface image is mapped to cubical Block.
Then it is converted using Snyder, maps to cubical piece of each vertex and be divided into multiple areas by cubical piece Domain.
Wherein it should be noted that the present embodiment described in step S305 can be with the execution of iteration, that is to say, that into pole Limit segmentation, i.e., the image-region that each image-region marks off only there are one before pixel, by image block it is continuous further Ground Split maps, to realize the minimum of mapping area.
Such as after the present embodiment S305 is finished, again by by the image-region that S305 is obtained be used as image block into One step is partitioned into the image-region more refined, and using cubical region as the image-region of block, again corresponding refinement, draws Subregion.It for the number of the execution of S305, can be adjusted according to practical computing capability and image quality requirements, the present invention The iterations of S305 are not limited.
Step S306, completion image is built on the surface of all spherical surface images there are mapping relations and cubical face After the mapping relations in region and region, surface and cubical face for all spherical surface images there are mapping relations, by ball The pixel projection of each image-region is to there are polyhedral regions in the block of mapping relations in the image block of face image.
Each image pixel of image-region and cubical face in surface in the present embodiment by calculating spherical surface image The mapping relations of middle region relative position, image pixel in image-region is projected to each region.That is, the present embodiment The middle projection that each pixel inside each net region is obtained using Snyder transformation.It is and every in the case of extreme segmentation There are one during pixel, the throwing of unique pixel inside region is directly only obtained by Snyder transformation for the inside of a image-region Shadow.
Embodiment 4
The method of a cubical image for 1-3 acquisitions or video image back projection to spherical surface will be performed as described above in the present embodiment Flow it is as follows：
Step S401, sphere is established, the surface area of cubical surface area and sphere spherical surface wherein where cube graph picture It is identical.By cube graph picture to spherical projection, spherical surface is divided into multiple surfaces.
Specific segmentation spherical surface further includes step S4011, is set to cube graph picture in the ball of spherical surface place, and with the centre of sphere Centered on point.
Step S4012, each vertex cubical in cube graph picture and seamed edge are mapped into spherical surface.
Spherical surface is divided into multiple surfaces by each vertex and seamed edge for step S4013, mapping to spherical surface.
Step S402, it is converted, is established respectively between each face of cube graph picture and each surface of spherical surface by Snyder Mapping relations.
For example, by obtaining mapping relations via Schneider transformation calculations as shown in Fig. 2, reference numeral 301 is signified [A, B, C, D] be face on cubic projection to spherical surface, reference numeral 303 signified [A', B', C', D'] is that cube is thrown Face on shadow to spherical surface.Specific vertex mapping relations are described referring to step 102 in embodiment 1, and details are not described herein again.
Step S403, respectively by cube graph as the pixel projection in each face is to there are the tables of the spherical surface of mapping relations Face.So as to complete conversion of the cube graph picture to the panoramic picture of spherical surface.
Embodiment 5
The cubical image of the present embodiment or the method for video image back projection are in the base of the back projection method of embodiment 4 On plinth, face further segmentation and the mapping projections, and then the distribution for further improving pixel is equal of spherical surface are projected to cube Even property.Specific method flow is as follows：
Step S501, sphere is established, the surface area of cubical surface area and sphere spherical surface wherein where cube graph picture It is identical.By cube graph picture to spherical projection, spherical surface is divided into multiple surfaces.
Specific segmentation spherical surface further includes step S5011, is set to cube graph picture in the ball of spherical surface place, and with the centre of sphere Centered on point.
Step S5012, each vertex cubical in cube graph picture and seamed edge are mapped into spherical surface.
Spherical surface is divided into multiple surfaces by each vertex and seamed edge for step S5013, mapping to spherical surface.
Step S502, it is converted, is established respectively between each face of cube graph picture and each surface of spherical surface by Snyder Mapping relations.Identical with the step S402 of embodiment, which is not described herein again.
Step S503, to every a pair, there are the following steps of surface execution in the face and spherical surface of the cube graph picture of mapping relations Suddenly：
S5031, the face of cube graph picture is evenly divided into multiple images block.
S5032, cube graph picture face on each vertex of each image block map to the surface of spherical surface.
S5033, converted by Snyder, map to the surface of spherical surface each vertex the surface of spherical surface is divided into it is multiple Block.
For example, each surface in the cube of the selection shown in reference numeral 201 is suitably divided.Such as Fig. 3 Shown in middle reference numeral 202, by way of drawing grid lines shown in reference numeral 203, cubical face is marked off several points Cut region.In the present embodiment, grid lines is all straight line, and each mesh segmentation region is the square of area equalization, attached Grid S shown in icon note 204, [A, B, C, D] is four vertex of grid S.
It is converted by Snyder, establishes cube and divide vertex to the mapping relations on spherical surface segmentation vertex.I.e. by attached drawing mark All grid vertexes shown in note 202 are all converted by Snyder on Algorithm mapping to spherical surface.By taking grid S as an example, need by [A, B, C, D] four points become scaling method by Snyder and corresponding [A', B', C', D'] coordinate position are calculated on spherical surface, i.e., Calculate [A, B, C, D] → [A', B', C', D'].
According to the mapping relations, the segmentation of cubical face is exactly projected to spherical surface by the projection of cube to spherical surface.Pass through The mapping relations of above-mentioned [A, B, C, D] → [A', B', C', D'] are established, the coordinate mapping relations of S → S' can be established.
Step S504, completion image is built on the surface in the face of all cube graph pictures there are mapping relations and spherical surface After the mapping relations of block and block, for all there are the face of cube graph picture of mapping relations and the surface of spherical surface, by cube The pixel projection of each image block is to there are the blocks in the surface of the spherical surface of mapping relations in the face of image.Cube is thrown to spherical surface Shadow, being exactly will be on the pixel projection to spherical surface in polyhedron S faces in S'.
Specifically, according to block in the vertex of image block in the surface of spherical surface image and cubical face in the present embodiment The mapping relations on vertex are projected image pixel in image block or image-region to block or region by pixel value difference algorithm.
Embodiment 6
The cubical image of the present embodiment or the method for video image back projection are in the base of the back projection method of embodiment 5 On plinth, simultaneously mapping projections are divided, and then improve the distributing homogeneity of pixel in the face that spherical surface is projected to cube again.Specifically Method flow is as follows：
Step S601, sphere is established, the surface area of cubical surface area and sphere spherical surface wherein where cube graph picture It is identical.By cube graph picture to spherical projection, spherical surface is divided into multiple surfaces.
Specific segmentation spherical surface further includes step S6011, is set to cube graph picture in the ball of spherical surface place, and with the centre of sphere Centered on point.
Step S6012, each vertex cubical in cube graph picture and seamed edge are mapped into spherical surface.
Spherical surface is divided into multiple surfaces by each vertex and seamed edge for step S6013, mapping to spherical surface.
Step S602, it is converted, is established respectively between each face of cube graph picture and each surface of spherical surface by Snyder Mapping relations.Identical with the step S402 of embodiment, which is not described herein again.
Step S603, to every a pair, there are the following steps of surface execution in the face and spherical surface of the cube graph picture of mapping relations Suddenly：
S6031, the face of cube graph picture is evenly divided into multiple images block.
S6032, cube graph picture face on each vertex of each image block map to the surface of spherical surface.
S6033, converted by Snyder, map to the surface of spherical surface each vertex the surface of spherical surface is divided into it is multiple Block.
Step S604, mapping is split to each image block again, specifically comprised the following steps：
S6041, image block is divided into multiple images region
S6042, it is converted by Snyder, establishes each figure with the image block of polyhedron image on the block of spherical surface respectively As the region there are mapping relations between region
Specifically, each vertex of each image-region of the image block of cube graph picture is mapped to the surface of spherical surface In block.
It is converted by Snyder, the block of spherical surface is divided into multiple areas by each vertex of the block mapped in the surface of spherical surface Domain.
Wherein it should be noted that the present embodiment described in step S604 can be with the execution of iteration, that is to say, that into pole Limit segmentation, i.e., the image-region that each image-region marks off only there are one before pixel, by image block it is continuous further Ground Split maps, to realize the minimum of mapping area.
Such as after the present embodiment S604 is finished, again by by the image-region that S604 is obtained be used as image block into One step is partitioned into the image-region more refined, and using the region of spherical surface as the image-region of block, again corresponding refinement, divides Region.It for the number of the execution of S604, can be adjusted according to practical computing capability and image quality requirements, the present invention is simultaneously The iterations of S604 are not limited.
Step S605, completion image is built on the surface in the face of all cube graph pictures there are mapping relations and spherical surface After the mapping relations in region and region, the face on surface and cube graph picture for all spherical surfaces there are mapping relations will be vertical The pixel projection of each image-region is to there are the regions in the block of the spherical surface of mapping relations in the image block of cube image
Each image pixel of image-region and area in spherical surface in surface in the present embodiment by calculating cube graph picture The mapping relations of domain relative position project image pixel in image-region to each region.That is, it is adopted in the present embodiment The projection of each pixel inside each net region is obtained with Snyder transformation.And each scheme in the case of extreme segmentation As the inside in region is only there are one during pixel, the uniquely projection of pixel is directly obtained inside region by Snyder transformation.
It is the contrast effect of embodiment 2 and the pixel distribution of prior art projection, wherein 401 institute of reference numeral shown in Fig. 4 Show be cylindrical surface projection after spherical surface pixel distribution density schematic diagram, reference numeral 402 is the spherical surface pixel after cubic projection Distribution density schematic diagram, reference numeral 403 are the spherical surface pixel distribution density schematic diagrames after the projecting method projection of embodiment 2, The projection scheme that visible embodiment 2 proposes from subjective vision effect can realize being uniformly distributed for pixel on spherical surface, the uniformity Remote super existing scheme.
In addition from objective data evaluation criterion, using spherical surface equator as starting point, with π/14 for step-length elevation angle statistical chart 4 In picture element density in the range of corresponding spherical surface.Chart shown in reference numeral 404 is under the projection of cylindrical surface in different elevation coverages Picture element density, chart shown in reference numeral 405 are the picture element density in different elevation coverages, reference numeral under cubic projection Chart shown in 406 be embodiment 2 projecting method under picture element density in different elevation coverages.It can be seen that it is evaluated in objective data On, the spherical surface pixel uniformity of embodiment 2 is better than the projecting method being currently known.
In addition, the assessment formula of J.Cui propositions can be used using the spherical surface point set uniformity：
Result of calculation D is lower, and the uniformity for representing point set is higher, and different pixels are given in Fig. 4 and are counted out corresponding calculating As a result.The spherical surface pixel uniformity of equally visible embodiment 2 is better than the projecting method being currently known.
Embodiment 7
The present embodiment proposes the image or process for video image projection applied to positive tetrahedron model, and based on the projection side Method further divides in the face of spherical projection to positive tetrahedron and mapping projections, and then further improves being evenly distributed for pixel Property.
Step S701, positive tetrahedron is established, by positive tetrahedron spherical projection, spherical surface image is divided into four surfaces.
Specific segmentation spherical surface image further includes step S7011, overlaps at positive tetrahedron center with sphere center position.
S7012, the surface that each vertex of positive tetrahedron and seamed edge are mapped to ball by perspective projection.
Spherical surface image is divided into multiple surfaces 506 by S7013, each vertex for mapping to spherical surface 505 and seamed edge.
S702, it is converted, is established respectively between each surface of spherical surface image and each face of positive tetrahedron by Snyder Mapping relations.
For example, mapping relations are obtained via Snyder transformation calculations as shown in figure 5, reference numeral 501 is referred to Positive tetrahedron, the signified face of reference numeral 502 correspond to the face 506 being projected on spherical surface 505.
Specific vertex mapping step is as follows：
The radius of spherical surface is wherein set as R, the positive tetrahedron model inscribe radius of a ball is r.As shown in figure 5, for face 506 Quadrangle segmentation may be used in partitioning scheme.
Specifically, each face of positive tetrahedron is divided into quadrilateral units according to symmetry, such as reference numeral 503 Mode divisional plane 502, the corresponding partitioning scheme that reference numeral 507 is just constituted to spherical surface.
The quadrilateral units S ' marked off as indicated by reference numeral 504, then corresponding surface is also divided into spherical surface four on spherical surface Side shape unit, the S as shown in reference numeral 508.
In addition another realization method of the present embodiment is divided using triangle, such as Fig. 6 of the division based on positive tetrahedron It is shown, each face on positive tetrahedron is divided into 6 triangular elements, then the corresponding surface on spherical surface is also divided into 6 Spherical triangle unit.As shown in reference numeral 601, face 601 is divided into 6 spherical triangles by tri- sides of AM, BN, CL Unit；Pass through three A ' M ', B ' N ', C ' L ' sides as shown in reference numeral 603, face 603 is divided into 6 triangular elements.Its The triangle 602 that each triangular element of middle positive tetrahedron is formed with AMI in the correspondence such as Fig. 6 of spherical triangle unit Correspondence between the triangle 604 of A ' M ' I ' compositions.Wherein, based on there are the triangular element of mapping relations and spherical surfaces Snyder transformation between triangular element is just repeated no more here with consistent described in embodiment 1.
S703, it after obtaining the mapping relations one by one put between positive tetrahedron surface and spherical surface, is incited somebody to action based on the mapping relations Spherical projection further divides to the face of positive tetrahedron and mapping projections, and then further improves the distributing homogeneity of pixel.
S7031, by the surface segmentation of spherical surface image be multiple images block.
S7032, it is converted by Snyder, establishes each image with the surface of spherical surface image on polyhedral face respectively There are the blocks of mapping relations between block.
Specific segmentation step is as follows：
It is converted by Snyder, polyhedral face is divided into multiple pieces by each vertex for mapping to polyhedral face.
For example, each sublist face to spherical surface is carried out according to each face of positive tetrahedron after suitably dividing.To each of spherical surface Spherical surface is marked off several cut zone by a sub- surface by way of drawing grid lines.S shown in reference numeral 508 is to divide The mesh segmentation region gone out, what [A, B, C, D] was characterized is four vertex of grid S.
It is converted by Snyder, establishes spherical surface and divide vertex to the mapping relations on positive tetrahedron segmentation vertex.I.e. by attached drawing All grid vertexes all convert Algorithm mapping to the face of positive tetrahedron 501 by Snyder in sublist face 506 shown in label 505 On 502.By taking grid S as an example, need [A, B, C, D] four points becoming scaling method in the face of positive tetrahedron 501 502 by Snyder On calculate corresponding [A', B', C', D'] coordinate position, i.e., grid S' shown in reference numeral 504, that is, calculate [A, B, C, D] → [A', B', C', D'].
According to the mapping relations, spherical surface to positive tetrahedron projects, and is exactly that spherical surface shown in reference numeral 507 is divided grid lines The segmentation grid lines of positive tetrahedron shown in reference numeral 503 is projected to, and then establishes above-mentioned [A, B, C, D] → [A', B', C', D'] Mapping relations, that is, establish the coordinate mapping relations of S → S'.
In addition, for positive tetrahedron, the unfolded surface of positive tetrahedron can also be become parallelogram, referring to Fig. 7, such as Shown in reference numeral 605, wherein the diamond shape being made of two neighboring face (such as face ACB knead dough BCD), and junction is divided The face cut is overlapped with two faces.Compared to other partitioning schemes, which can cause the distribution in sublist face more Uniformly, shape is more consistent.
Hereafter further can also parallelogram image be converted to so that regard by rectangle diagram picture by affine transformation Frequency image is suitable for existing coding techniques, as designated by reference numeral 606.Other modes can also be used, such as can will be put down Row quadrangle polishing is rectangle, and as shown in reference numeral 607, in the rectangle diagram picture after polishing, polishing part can be filled out Fill be pure color without influencing code efficiency rather than polishing part, i.e., former parallelogram-shaped segment still fills with former parallelogram Image.
It can by the description of the specific embodiment of the projecting method and back projection method of above panoramic video and picture Know, those skilled in the art can be understood that the projection of the panoramic video and picture of the application and back projection can be by Software adds the mode of required general hardware platform to realize.Based on such understanding, the technical solution of the application substantially or Person says can embody the part that the prior art contributes in the form of software product, and the computer software product can To be stored in storage medium, such as ROM/RAM (read-only memory/random access memory), magnetic disc, CD, including several Instruction, which is used, so that computer equipment (can be personal computer, server or the network equipment etc.) performs the application Method described in certain parts of embodiment or embodiment.
The application can be used in numerous general or special purpose computing system environments or configuration.Such as:Personal computer, service Device computer, handheld device or portable device, laptop device, multicomputer system, the system based on microprocessor, top set Box, programmable consumer-elcetronics devices, network PC (PC), minicomputer, mainframe computer, including any of the above system Distributed computing environment of system or equipment etc..
The application can be described in the general context of computer executable instructions, such as program Module.Usually, program module includes routines performing specific tasks or implementing specific abstract data types, program, object, group Part, data structure etc..The application can also be put into practice in a distributed computing environment, in these distributed computing environment, by Task is performed and connected remote processing devices by communication network.In a distributed computing environment, program module can be with In the local and remote computer storage media including storage device.
Although specific embodiments of the present invention have been described above, it will be appreciated by those of skill in the art that these It is merely illustrative of, protection scope of the present invention is defined by the appended claims.Those skilled in the art is not carrying on the back Under the premise of from the principle and substance of the present invention, many changes and modifications may be made, but these are changed Protection scope of the present invention is each fallen with modification.

Claims (34)

  1. A kind of 1. projecting method of panoramic video and picture, which is characterized in that the projecting method packet of the panoramic video and picture Include following steps：
    By polyhedron to spherical projection, spherical surface image is divided into multiple surfaces；
    By mapping transformation, the mapping relations between each surface of spherical surface image and polyhedral each face are established respectively；
    Respectively by the pixel projection in each surface of spherical surface image to there are polyhedral faces of mapping relations.
  2. 2. the projecting method of panoramic video as described in claim 1 and picture, which is characterized in that by polyhedron to spherical surface Projection, includes the following steps during spherical surface image is divided into multiple surfaces：
    Where polyhedron is set to spherical surface image in sphere, and the point centered on the centre of sphere；
    Polyhedral each vertex and seamed edge are mapped into spherical surface image；
    Spherical surface image is divided into multiple surfaces by each vertex and seamed edge for mapping to spherical surface image.
  3. 3. the projecting method of panoramic video as described in claim 1 and picture, which is characterized in that by polyhedron to spherical surface Projection, the Cheng Qian excessively that spherical surface image is divided into multiple surfaces are further comprising the steps of：
    Polyhedron is established, wherein polyhedral surface area is identical with the surface area of sphere where spherical surface image.
  4. 4. the projecting method of panoramic video as described in claim 1 and picture, which is characterized in that
    Include during pixel projection to polyhedral face there are mapping relations in each surface by spherical surface image Following steps：
    To every a pair, there are the surfaces of the spherical surface image of mapping relations and polyhedral face execution following steps：
    By the surface segmentation of spherical surface image into multiple images block；
    By mapping transformation, establish to exist between each image block on the surface of spherical surface image respectively on polyhedral face and reflect Penetrate the block of relationship；
    The mapping for completing image block and block is built in the surface of all spherical surface images there are mapping relations and polyhedral face After relationship, surface and polyhedral face to all spherical surface images there are mapping relations perform following steps：
    By the pixel projection of image block each in the surface of spherical surface image to the block in polyhedral face there are mapping relations.
  5. 5. the projecting method of panoramic video as claimed in claim 4 and picture, which is characterized in that established on polyhedral face Between each image block on the surface of spherical surface image there are the block of mapping relations during include the following steps：
    Each vertex of image block each in the surface of spherical surface image is mapped into polyhedral face；
    By mapping transformation, polyhedral face is divided into multiple pieces by each vertex for mapping to polyhedral face.
  6. 6. the projecting method of panoramic video as claimed in claim 4 and picture, which is characterized in that established on polyhedral face There are iteration after the process of the block of mapping relations between each image block on the surface of spherical surface image to perform following steps：
    By image block into multiple images region；
    By mapping transformation, establish between each image-region of the image block of spherical surface image deposit on polyhedral piece respectively In the region of mapping relations；
    It is built in the surface of all spherical surface images there are mapping relations and polyhedral face and completes image-region and region After mapping relations, surface and polyhedral face to all spherical surface images there are mapping relations perform following steps：
    By the pixel projection of image-region each in the image block of spherical surface image to there are the polyhedral in the block of mapping relations Region.
  7. 7. the projecting method of panoramic video as claimed in claim 6 and picture, which is characterized in that established on polyhedral piece Between each image-region of the image block of spherical surface image there are the region of mapping relations during include the following steps：
    Each vertex of each image-region of the image block of spherical surface image is mapped to polyhedral piece；
    By mapping transformation, map to polyhedral piece of each vertex and be divided into multiple regions by polyhedral piece.
  8. 8. the projecting method of the panoramic video and picture as described in any one of claim 4-7, which is characterized in that by spherical surface In the surface of image the pixel projection of each image block or image-region to the block in polyhedral face there are mapping relations or Include the following steps during region：
    According to the vertex in block or region in the vertex of image block or image-region in the surface of spherical surface image and polyhedral face Mapping relations are projected image pixel in image block or image-region to block or region by pixel value difference algorithm；Or
    Calculate each image pixel of image block or image-region and block or region in polyhedral face in the surface of spherical surface image The mapping relations of relative position project image pixel in image block or image-region to block or region.
  9. 9. the projecting method of the panoramic video and picture as described in any one of claim 4-7, which is characterized in that by spherical surface During the surface of image or image block are divided into multiple images block or image-region, by the surface of spherical surface image or image block It is evenly divided into multiple images block or image-region.
  10. 10. the projecting method of panoramic video as claimed in claim 9 and picture, which is characterized in that built on polyhedral face There are the block of mapping relations or the processes in region between each image block or each image-region on vertical and spherical surface image surface In, the area equation of the block or region of each image block or image-region mapping.
  11. 11. the projecting method of the panoramic video and picture as described in any one of claim 4-7, which is characterized in that each piece Or region is quadrangle.
  12. 12. the projecting method of panoramic video as claimed in claim 11 and picture, which is characterized in that each piece or region is just It is rectangular.
  13. 13. the projecting method of the panoramic video and picture as described in any one of claim 1-7, which is characterized in that described to reflect Penetrate be transformed to Snyder transformation, RD transformation, Cube transformation, Unicube transformation, IsoCube transformation, perspective projection transformation or Lambert authalic projections convert.
  14. 14. the projecting method of the panoramic video and picture as described in any one of claim 1-7, which is characterized in that described more Face body is regular polygon.
  15. 15. the projecting method of panoramic video as claimed in claim 14 and picture, which is characterized in that the polyhedron is positive four Face body, cube or regular dodecahedron.
  16. 16. the projecting method of the panoramic video and picture as described in any one of claim 1-7, which is characterized in that described each A surface includes multiple images pixel.
  17. 17. the projecting method of the panoramic video and picture as described in any one of claim 4-7, which is characterized in that described each A image block includes at least one image pixel or each image-region includes at least one image pixel.
  18. A kind of 18. back projection method of panoramic video and picture, which is characterized in that the back projection side of the panoramic video and picture Method includes the following steps：
    By polyhedron image to spherical projection, spherical surface is divided into multiple surfaces；
    By mapping transformation, the mapping relations between each face of polyhedron image and each surface of spherical surface are established respectively；
    Respectively by the pixel projection in each face of polyhedron image to there are the surfaces of the spherical surface of mapping relations.
  19. 19. the back projection method of panoramic video as claimed in claim 18 and picture, which is characterized in that described to pass through polyhedron Image includes the following steps to spherical projection during spherical surface is divided into multiple surfaces：
    Where polyhedron image is set to spherical surface in ball, and the point centered on the centre of sphere；
    Each vertex polyhedral in polyhedron image and seamed edge are mapped into spherical surface；
    Spherical surface is divided into multiple surfaces by each vertex and seamed edge for mapping to spherical surface.
  20. 20. the back projection method of panoramic video as claimed in claim 18 and picture, which is characterized in that by polyhedron to Spherical projection, the Cheng Qian excessively that spherical surface is divided into multiple surfaces are further comprising the steps of：
    Sphere is established, wherein polyhedral surface area is identical with the surface area of sphere spherical surface where polyhedron image.
  21. 21. the back projection method of panoramic video as claimed in claim 18 and picture, which is characterized in that by polyhedron image Pixel projection in each face is to including the following steps during there are the surface of the spherical surface of mapping relations：
    To every a pair, there are the faces of the polyhedron image of mapping relations and the surface execution following steps of spherical surface：
    The face of polyhedron image is evenly divided into multiple images block；
    By mapping transformation, exist between each image block in the face of foundation and polyhedron image on the surface of spherical surface reflect respectively Penetrate the block of relationship；
    The mapping for completing image block and block is built on the face of all polyhedron images there are mapping relations and the surface of spherical surface After relationship, the surface in face and spherical surface to all polyhedron images there are mapping relations performs following steps：
    By the pixel projection of image block each in the face of polyhedron image to there are the blocks in the surface of the spherical surface of mapping relations.
  22. 22. the back projection method of panoramic video as claimed in claim 21 and picture, which is characterized in that on the surface of spherical surface Include the following steps during establishing the block between each image block of polyhedron image there are mapping relations：
    Each vertex of image block each on the face of polyhedron image is mapped to the surface of spherical surface；
    By mapping transformation, the surface of spherical surface is divided into multiple pieces by each vertex for mapping to the surface of spherical surface.
  23. 23. the back projection method of panoramic video as claimed in claim 21 and picture, which is characterized in that on the surface of spherical surface There are iteration after the process of the block of mapping relations between foundation and each image block in the face of polyhedron image to perform following steps：
    Image block is divided into multiple images region；
    By mapping transformation, establish between each image-region of the image block of polyhedron image deposit on the block of spherical surface respectively In the region of mapping relations；
    It is built on the face of all polyhedron images there are mapping relations and the surface of spherical surface and completes image-region and region After mapping relations, the face on surface and polyhedron image to all spherical surfaces there are mapping relations performs following steps：
    By the pixel projection of image-region each in the image block of polyhedron image to there are the in the block of the spherical surface of mapping relations Region.
  24. 24. the back projection method of panoramic video as claimed in claim 23 and picture, which is characterized in that built on the block of spherical surface It is vertical between each image-region of the image block of polyhedron image there are the region of mapping relations during including following step Suddenly：
    Each vertex of each image-region of the image block of polyhedron image is mapped to the block in the surface of spherical surface；
    By mapping transformation, the block of spherical surface is divided into multiple regions by each vertex of the block mapped in the surface of spherical surface.
  25. 25. the back projection method of the panoramic video and picture as described in any one of claim 21-24, which is characterized in that By the pixel projection of image block each in the face of polyhedron image or image-region in there are the surface of the spherical surface of mapping relations Block or region during include the following steps：
    According to the vertex in block or region in the surface on the vertex and spherical surface of image block or image-region in the face of polyhedron image Mapping relations are projected image pixel in image block or image-region to block or region by pixel value difference algorithm；Or
    Calculate polyhedron image face in image block or image-region each image pixel and spherical surface surface in block or region The mapping relations of relative position project image pixel in image block or image-region to block or region.
  26. 26. the back projection method of the panoramic video and picture as described in any one of claim 21-24, which is characterized in that During the face of polyhedron image or image block are divided into multiple images block or image-region, by the face of polyhedron image or Image block is evenly divided into multiple images block or image-region.
  27. 27. the back projection method of panoramic video as claimed in claim 26 and picture, which is characterized in that on the surface of spherical surface There are the block of mapping relations or the mistakes in region between foundation and each image block in the face of polyhedron image or each image-region The area equation of Cheng Zhong, each image block or image-region.
  28. 28. the back projection method of the panoramic video and picture as described in any one of claim 21-24, which is characterized in that each A image block or image-region are quadrangle.
  29. 29. the back projection method of panoramic video as claimed in claim 28 and picture, which is characterized in that each image block or figure As region is square.
  30. 30. the back projection method of the panoramic video and picture as described in any one of claim 17-24, which is characterized in that institute State mapping transformation for Snyder transformation, RD transformation, Cube transformation, Unicube transformation, IsoCube transformation, perspective projection transformation or Lambert authalic projections convert.
  31. 31. the back projection method of the panoramic video and picture as described in any one of claim 17-24, which is characterized in that institute Polyhedron is stated as regular polygon.
  32. 32. the back projection method of panoramic video as claimed in claim 31 and picture, which is characterized in that the polyhedron is just Tetrahedron, cube or regular dodecahedron.
  33. 33. the back projection method of the panoramic video and picture as described in any one of claim 17-24, which is characterized in that institute State polyhedral each bread pixel containing multiple images.
  34. 34. the back projection method of the panoramic video and picture as described in any one of claim 21-24, which is characterized in that institute It states each image block and includes at least one image pixel including at least one image pixel or each image-region.
CN201611209451.1A 2016-12-23 2016-12-23 The projecting method and back projection method of panoramic video and picture Pending CN108243333A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201611209451.1A CN108243333A (en) 2016-12-23 2016-12-23 The projecting method and back projection method of panoramic video and picture

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201611209451.1A CN108243333A (en) 2016-12-23 2016-12-23 The projecting method and back projection method of panoramic video and picture

Publications (1)

Publication Number Publication Date
CN108243333A true CN108243333A (en) 2018-07-03

Family

ID=62704396

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201611209451.1A Pending CN108243333A (en) 2016-12-23 2016-12-23 The projecting method and back projection method of panoramic video and picture

Country Status (1)

Country Link
CN (1) CN108243333A (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109559675A (en) * 2018-11-02 2019-04-02 东南大学 A kind of warp direction addressing display methods based on panorama spherical surface display
CN110889048A (en) * 2018-08-20 2020-03-17 高德信息技术有限公司 Map data query method, system, terminal and server
CN112770095A (en) * 2021-01-28 2021-05-07 广州方硅信息技术有限公司 Panoramic projection method and device and electronic equipment
CN113163184A (en) * 2021-03-25 2021-07-23 复旦大学 Projection algorithm for VR self-adaptive transmission
CN116760965A (en) * 2023-08-14 2023-09-15 腾讯科技（深圳）有限公司 Panoramic video encoding method, device, computer equipment and storage medium

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110889048A (en) * 2018-08-20 2020-03-17 高德信息技术有限公司 Map data query method, system, terminal and server
CN110889048B (en) * 2018-08-20 2022-09-09 阿里巴巴（中国）有限公司 Map data query method, system, terminal and server
CN109559675A (en) * 2018-11-02 2019-04-02 东南大学 A kind of warp direction addressing display methods based on panorama spherical surface display
CN109559675B (en) * 2018-11-02 2020-07-24 东南大学 Warp direction addressing display method based on panoramic spherical display
CN112770095A (en) * 2021-01-28 2021-05-07 广州方硅信息技术有限公司 Panoramic projection method and device and electronic equipment
CN113163184A (en) * 2021-03-25 2021-07-23 复旦大学 Projection algorithm for VR self-adaptive transmission
CN113163184B (en) * 2021-03-25 2022-11-18 复旦大学 Projection algorithm for VR self-adaptive transmission
CN116760965A (en) * 2023-08-14 2023-09-15 腾讯科技（深圳）有限公司 Panoramic video encoding method, device, computer equipment and storage medium
CN116760965B (en) * 2023-08-14 2023-12-22 腾讯科技（深圳）有限公司 Panoramic video encoding method, device, computer equipment and storage medium

Similar Documents

Publication Publication Date Title
CN108243333A (en) The projecting method and back projection method of panoramic video and picture
CN106296783B (en) A kind of space representation method of combination space overall situation 3D view and panoramic pictures
US9776086B2 (en) Method of transforming an image file
EP3534336B1 (en) Panoramic image generating method and apparatus
US9626790B1 (en) View-dependent textures for interactive geographic information system
US6683608B2 (en) Seaming polygonal projections from subhemispherical imagery
CN104183016B (en) A kind of construction method of quick 2.5 dimension building model
CN102054290B (en) Construction method of panoramic/realistic hybrid reality platform
CN107622474A (en) Panoramic video mapping method based on main view point
CN107506038A (en) A kind of three-dimensional earth exchange method based on mobile terminal
CN108876910A (en) A kind of three-dimensional modeling method, device, equipment and computer readable storage medium
CN108564527A (en) The method and device of the completion of panorama sketch content and reparation based on neural network
CN107993282A (en) One kind can dynamically measure live-action map production method
CN106780317A (en) panoramic projection method and device
Livny et al. A GPU persistent grid mapping for terrain rendering
US5694531A (en) Method and apparatus for simulating movement in multidimensional space with polygonal projections
Dachsbacher et al. Rendering Procedural Terrain by Geometry Image Warping.
Beck Real-time visualization of big 3D city models
CN109328462A (en) A kind of method and device for stream video content
Tian et al. Occlusion handling using moving volume and ray casting techniques for augmented reality systems
CN108230430B (en) Cloud layer mask image processing method and device
CN108616747A (en) The playback method of panoramic video and picture
Hladky et al. Snakebinning: Efficient temporally coherent triangle packing for shading streaming
Kada et al. Real-time visualisation of urban landscapes using open-source software
AU2020101803A4 (en) Method for Optimizing the Coding Block-Level Lagrange Multiplier of Longitude-Latitude Image

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
WD01 Invention patent application deemed withdrawn after publication
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20180703