CN102780834A - Semi-cylindrical panorama upwrapping method of pano-eye image - Google Patents
Semi-cylindrical panorama upwrapping method of pano-eye image Download PDFInfo
- Publication number
- CN102780834A CN102780834A CN2011101205467A CN201110120546A CN102780834A CN 102780834 A CN102780834 A CN 102780834A CN 2011101205467 A CN2011101205467 A CN 2011101205467A CN 201110120546 A CN201110120546 A CN 201110120546A CN 102780834 A CN102780834 A CN 102780834A
- Authority
- CN
- China
- Prior art keywords
- image
- semicylinder
- eye lens
- semi
- eye
- 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
Landscapes
- Studio Devices (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a semi-cylindrical panorama upwrapping method of a pano-eye image. The semi-cylindrical panorama upwrapping method comprises the following steps of: at first, according to projection relation of imaging any one point in a semi-cylindrical space with an angle range of 0-180 degrees to an omni-directional image through a pano-eye lens, mapping a target pixel point in the semi-cylindrical space to an originally collected pano-eye lens image; and then obtaining a semi-cylindrical panorama upwrapping image of the pano-eye lens image according to coordinate mapping relation from any one point in the semi-cylindrical space to a semi-cylindrical panorama upwrapping plane. By utilizing the semi-cylindrical panorama upwrapping method provided by the invention, the problems of severe distortion in image correction, image information loss and the like, caused by using the traditional spherical projection correction method of a fisheye image and the traditional plane perspective projection correction method, can be avoided, so that the image quality and visual effect of the semi-cylindrical panorama upwrapping image of the pano-eye image can be greatly improved.
Description
Technical field:
The present invention relates to the omnidirectional vision field, refer in particular to the semicylinder panoramic expansion method of a kind of eye lens image (Pano-Eye Lens Image).
Background technology:
Panoramic picture/video relies on the comprehensive visual angle of 360 degree, can cover the advantage of scene on a large scale, has obtained research and application more and more widely in fields such as panoramic video monitoring, robot visual guidance, virtual real-image space structures.The eye camera lens is a kind of fish-eye bugeye lens that is similar to, and the imaging effect of 180 ° even bigger field range can be provided.Panorama picture of fisheye lens face zones of different, its resolution is different, and the closer to picture centre, resolution is high more, and detailed information is many more, slip chart inconocenter more, resolution is low more, and detailed information is few more, is out of shape seriously more, and tangible barrel-shaped distortion (distortion) is arranged.And eye camera lens (Pano-Eye Lens) makes picture centre and edge resolution difference reduce through special optical design, thereby has reduced distortion, has improved the panorama observing effect.Also not to the method for correcting image of eye camera lens, common correcting fisheye image method mainly contains following several kinds at present:
The spherical projection model is a kind of simple effective method, regards the panorama picture of fisheye lens face as a sphere.But this method need be known the optical centre of fish eye images and the radius of conversion sphere in advance.Therefore existing method only is applicable to the fish eye images with border circular areas." a kind of fish eye lens bearing calibration based on spherical perspective projection constraint " that such classical algorithm such as Ying Xianghua and Hu Zhanyi propose (" Chinese journal of computers ", 2003, Vol.26, No.12).
Parabolic imaging model more complicated is regarded the panorama picture of fisheye lens face as a parabola, when the restoration scenario degree of depth, can obtain accurate more effect.But too complicated during with this Model Calculation, generally be used to utilize the flake photo to recover the depth information technology.Relevant research can get by reference papers Jia Yun, Lv Hongjing, " the fish eye lens stereo-picture recovers the method for dense depth map " of Liu Wanchun (" Chinese journal of computers ", 2003, Vol.23, No.12).
The distortion of 2D fish eye images is proofreaied and correct, and this method does not relate to spatial point information, directly confirms corresponding points coordinate transform on deformation pattern and the image to be corrected, carries out the pixel grey scale interpolation then.These class methods include spherical coordinate location, multinomial coordinate transform, projective invariance and shine upon through polar radius and to proofread and correct the fish eye images distortion.The algorithm that launch the spherical coordinate location is put forward in paper " Techniques for Fisheye Lens Calibaration using a Minimal Number of Measurements " (Proc.of the SPIE Intelligent Robotics and Computer Vision Conference, 2000) by T.Nathan Mundhenk the earliest.
The distortion of 3D fish eye images is proofreaied and correct, and comprises two kinds of methods of projection conversion and fish eye lens demarcation.The projection transfer algorithm is the image that fish eye images is converted to perspective projection, is mapped to the two dimensional millet cake that the three-dimensional scenic projection constitutes to each image plane vegetarian refreshments on the fish eye images, according to the relation between image slices vegetarian refreshments and corresponding light three-dimensional vector, realizes proofreading and correct.Classic methods is seen paper " Generalized Camera Calibration Including Fish-Eye Lenses " (the International Journal of Computer Vision of D.B.Gennery; 2006), paper " Can We Consider Central Catadioptric Cameras and Fisheye Cameras within a Unified Imaging Model " (ECCV 2004) of X.H Ying etc. etc.The fish eye lens calibration algorithm is one type of method of accurately recovering; On the basis of setting up the fish eye lens distorted pattern; Consider the various distortion types of imaging; Like common radial deformation, centrifugal distortion, thin prism distortion etc., solve the fish eye lens internal and external parameter through experiment and target function, thereby reach accurate recovery fish eye images distortion." a kind of scaling method of panorama picture of fisheye lens stereo visual system " that this type of algorithm can must wait referring to Jia Yun (" Chinese journal of computers ", 2000, Vol.23, No.11) etc.
Yet, the part that still comes with some shortcomings of correcting fisheye image method in the past, as proofread and correct the image border, back and have comparatively serious deformation, image raw information lose complex algorithm or the like.The present invention is directed to the deficiency of traditional correcting fisheye image method, proposed a kind of semicylinder panoramic expansion method to the eye lens image.
Summary of the invention:
The technical problem that the present invention will solve is: proofread and correct this particular problem to the eye lens image; Problems such as the image rectification distortion for fear of traditional fish eye images spherical projection bearing calibration and plane perspective projection bearing calibration etc. causes is serious, image information is lost; A kind of semicylinder panoramic expansion method is proposed; Through calculating original eye lens image and proofreading and correct the coordinate mapping relations between the cylinder panoramic image of back, picture quality and visual effect after the eye lens image is proofreaied and correct have been improved.
For solving the problems of the technologies described above, the technical scheme that the present invention takes comprises the steps:
1. according to 0 ° of-180 ° of projection relation that semicylinder space any point is schemed to omnidirectional through the eye lens imaging, each the pixel p in the semicylinder space is mapped to the corresponding points p ' in the eye lens image;
2. the coordinate mapping relations of the corresponding points t of any point p in the semicylinder panoramic expansion plane in the derivation semicylinder space;
3. according to step 1 and step 2, deriving obtains the mapping relations of corresponding points p ' in arbitrary pixel t and the eye lens image on the semicylinder panoramic expansion plane;
4. travel through each pixel in the semicylinder panoramic expansion plane, repeating step 1-3 is until generating 0 ° of complete-180 ° of semicylinder panoramic picture.
Wherein the projection relation of scheming to omnidirectional through the eye lens imaging of the semicylinder space any point in the step 1 is the parameters according to the eye camera lens that is designed, and deriving through the optical imagery geometrical analysis obtains.The used eye camera lens of the present invention is different from the traditional standard fish eye lens and satisfies r basically
s=f θ mapping, but
Relation uses high-precision non-spherical element to reduce the abaxial effect that is extruded, and the f-θ design than standard has reduced distortion, so just makes the image border obtain more pixel.Wherein
Be the function of θ, satisfy
Plane aperture perspective imaging satisfies r
p=ftan θ, (shown in accompanying drawing 1 and accompanying drawing 2).
Step 2 is the coordinate mapping relations according to the corresponding points t of any point p in the transformation relation derivation semicylinder space of planar coordinate and 3 d space coordinate in the semicylinder panoramic expansion plane.
Description of drawings:
Fig. 1 is eye camera lens, traditional fish eye lens and common perspective camera light mapping relations comparison diagram;
Fig. 2 is light mapping relations and traditional standard fish eye lens and the satisfied light mapping relations comparison diagram of common lens that the used eye camera lens of the present invention satisfies;
Fig. 3 is the projection relation sketch map that any point is schemed to omnidirectional through the eye lens imaging in the semicylinder space;
Fig. 4 is the coordinate mapping relations sketch map of the corresponding points of any point in the semicylinder space in the semicylinder panoramic expansion plane;
Fig. 5 is eye lens image semicylinder panoramic expansion method actual effect figure.
Embodiment:
Embodiment 1: with the used eye lens imaging of the present invention is example, in conjunction with accompanying drawing the present invention is explained further details.
The first step:
The projection of eye lens imaging be meant with arbitrarily some mapping in the three dimensions through the eye lens imaging to the picture plane a bit.Shown in accompanying drawing 3, any point p (x in the semicylinder space
p, y
p, z
p) an incident ray pO sending meets at a p ' (x through aperture center O and imaging plane xO ' y
f, y
f), some p ' is the subpoint of spatial point p in the eye lens image.
In eye lens image plane F; Any point p ' in the image can use formula (1) to obtain, and wherein r is a p ' to eye lens image center O ' distance, θ is the angle of visual field; Be the angle of incident ray and optical axis, f is equivalent focal length (Effective Focal Length).
Angle of visual field θ through type (2) obtains:
Azimuth λ
p, i.e. the angle of incident ray and x axle, through type (3) obtains:
Therefore, any point p (x in the semicylinder space
p, y
p, z
p) and eye lens imaging plane xO ' y corresponding points p ' (x
f, y
f) mapping relations suc as formula shown in (4):
Second step:
Shown in accompanying drawing 4, any point t (x among the semicylinder panoramic expansion plane T
t, y
t) corresponding points p (x in the P of semicylinder space
p, y
p, z
p) the coordinate mapping relations, therefore accompanying drawing 4 (c) is the semicylinder cross sectional representation, and following relational expression is arranged:
x
p=r?cos(α) (5)
y
p=y
t (6)
z
p=r?sin(α) (7)
And have:
Wherein, W is the width of panoramic expansion figure.
At last, in conjunction with the result in two steps, just can derive obtains the mapping relations of corresponding points p ' in arbitrary pixel t and the eye lens image on the semicylinder panoramic expansion plane, reached the present invention the target that will realize.
The semicylinder panoramic expansion figure that is obtained by the semicylinder panoramic expansion method of above-mentioned eye lens image is shown in accompanying drawing 5, and picture quality is high, and it is little to proofread and correct back image fault distortion, and the image level direction does not have information dropout.
In fact the method that proposes among the present invention can be used for general flake or wide-angle lens imaging panoramic expansion method.Above embodiment only plays the effect of explaining technical scheme of the present invention, and protection range of the presently claimed invention is not limited to described implementation method of the foregoing description and practical implementation step.Therefore, only imaging model concrete in the foregoing description, formula and algorithm are simply replaced, but its flesh and blood still with the corresponding to technical scheme of the method for the invention, all should belong to protection scope of the present invention.
Claims (3)
1. the semicylinder panoramic expansion method of an eye lens image (Pano-Eye Image), its characteristic comprises the steps:
1) according to 0 ° of-180 ° of projection relation that semicylinder space any point is schemed to omnidirectional through the eye lens imaging, each the pixel p in the semicylinder space is mapped to the corresponding points p ' in the eye lens image;
2) the coordinate mapping relations of the corresponding points t of any point p in the semicylinder panoramic expansion plane in the derivation semicylinder space;
3) according to step 1 and step 2, deriving obtains the mapping relations of corresponding points p ' in arbitrary pixel t and the eye lens image on the semicylinder panoramic expansion plane;
4) each pixel in the traversal semicylinder panoramic expansion plane, repeating step 1)-3), until generating 0 ° of complete-180 ° of semicylinder panoramic expansion image.
2. according to the said eye lens image of claim 1 semicylinder panoramic expansion method; The projection relation that semicylinder space any point in the said step 1) is schemed to omnidirectional through the eye lens imaging is the parameters according to the eye camera lens that is designed, and obtains through optical imagery geometrical analysis derivation; The used eye camera lens of the present invention uses high-precision non-spherical element to reduce the abaxial effect that is extruded, and than the fish eye lens f-θ design of standard, distortion reduces, and makes the image border obtain more pixel.
3. according to the said eye lens image of claim 1 semicylinder panoramic expansion method, said step 2) in the semicylinder space in the coordinate mapping relations of the corresponding points t of any point p in the semicylinder panoramic expansion plane be to derive according to the transformation relation of planar coordinate and 3 d space coordinate.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2011101205467A CN102780834A (en) | 2011-05-11 | 2011-05-11 | Semi-cylindrical panorama upwrapping method of pano-eye image |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2011101205467A CN102780834A (en) | 2011-05-11 | 2011-05-11 | Semi-cylindrical panorama upwrapping method of pano-eye image |
Publications (1)
Publication Number | Publication Date |
---|---|
CN102780834A true CN102780834A (en) | 2012-11-14 |
Family
ID=47125582
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN2011101205467A Pending CN102780834A (en) | 2011-05-11 | 2011-05-11 | Semi-cylindrical panorama upwrapping method of pano-eye image |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN102780834A (en) |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103325110A (en) * | 2013-05-29 | 2013-09-25 | 山西绿色光电产业科学技术研究院(有限公司) | Panoramic image correction algorithm based on panoramic all-in-one speed dome camera |
CN105812759A (en) * | 2016-04-15 | 2016-07-27 | 杭州当虹科技有限公司 | Planar projection method and coding method of 360-degree panoramic video |
CN106993126A (en) * | 2016-05-11 | 2017-07-28 | 深圳市圆周率软件科技有限责任公司 | A kind of method and device that lens image is expanded into panoramic picture |
CN107845056A (en) * | 2016-09-21 | 2018-03-27 | 西安英诺视通信息技术有限公司 | Fish eye images panorama generation method based on cylinder model |
CN109816587A (en) * | 2017-11-20 | 2019-05-28 | 杭州海康威视数字技术股份有限公司 | A kind of fish eye images processing method, device and electronic equipment |
CN110930299A (en) * | 2019-12-06 | 2020-03-27 | 杭州视洞科技有限公司 | Circular fisheye video display scheme based on hemisphere expansion |
WO2021082264A1 (en) * | 2019-10-30 | 2021-05-06 | 歌尔股份有限公司 | Projection image automatic correction method and system based on binocular vision |
CN113971703A (en) * | 2020-07-22 | 2022-01-25 | 阿里巴巴集团控股有限公司 | Image processing method and device, electronic equipment and storage medium |
-
2011
- 2011-05-11 CN CN2011101205467A patent/CN102780834A/en active Pending
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103325110A (en) * | 2013-05-29 | 2013-09-25 | 山西绿色光电产业科学技术研究院(有限公司) | Panoramic image correction algorithm based on panoramic all-in-one speed dome camera |
CN103325110B (en) * | 2013-05-29 | 2016-06-01 | 湖南源信光电科技有限公司 | A kind of panoramic picture correction method based on panorama clipping the ball all-in-one |
CN105812759A (en) * | 2016-04-15 | 2016-07-27 | 杭州当虹科技有限公司 | Planar projection method and coding method of 360-degree panoramic video |
CN106993126A (en) * | 2016-05-11 | 2017-07-28 | 深圳市圆周率软件科技有限责任公司 | A kind of method and device that lens image is expanded into panoramic picture |
CN107845056A (en) * | 2016-09-21 | 2018-03-27 | 西安英诺视通信息技术有限公司 | Fish eye images panorama generation method based on cylinder model |
CN109816587A (en) * | 2017-11-20 | 2019-05-28 | 杭州海康威视数字技术股份有限公司 | A kind of fish eye images processing method, device and electronic equipment |
WO2021082264A1 (en) * | 2019-10-30 | 2021-05-06 | 歌尔股份有限公司 | Projection image automatic correction method and system based on binocular vision |
US11606542B2 (en) | 2019-10-30 | 2023-03-14 | Goertek Inc. | Projection image automatic correction method and system based on binocular vision |
CN110930299A (en) * | 2019-12-06 | 2020-03-27 | 杭州视洞科技有限公司 | Circular fisheye video display scheme based on hemisphere expansion |
CN113971703A (en) * | 2020-07-22 | 2022-01-25 | 阿里巴巴集团控股有限公司 | Image processing method and device, electronic equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102780834A (en) | Semi-cylindrical panorama upwrapping method of pano-eye image | |
CN106600546B (en) | Distortion correction method and system for ultra-wide-angle camera | |
CN110211043B (en) | Registration method based on grid optimization for panoramic image stitching | |
Won et al. | Omnimvs: End-to-end learning for omnidirectional stereo matching | |
WO2019100933A1 (en) | Method, device and system for three-dimensional measurement | |
CN103198524B (en) | A kind of three-dimensional reconstruction method for large-scale outdoor scene | |
US20190012804A1 (en) | Methods and apparatuses for panoramic image processing | |
AU2008283197B9 (en) | Methods of obtaining panoramic images using rotationally symmetric wide-angle lenses and devices thereof | |
CN108288292A (en) | A kind of three-dimensional rebuilding method, device and equipment | |
CN108734657B (en) | Image splicing method with parallax processing capability | |
CN106228538A (en) | Binocular vision indoor orientation method based on logo | |
CN107545586B (en) | Depth obtaining method and system based on light field polar line plane image local part | |
CN106534670B (en) | It is a kind of based on the panoramic video generation method for connecting firmly fish eye lens video camera group | |
CN104778656A (en) | Fisheye image correction method on basis of spherical perspective projection | |
CN105825470A (en) | Fisheye image correction method base on point cloud image | |
CN116958437A (en) | Multi-view reconstruction method and system integrating attention mechanism | |
CN103247020A (en) | Fisheye image spread method based on radial characteristics | |
CN113298947A (en) | Multi-source data fusion-based three-dimensional modeling method medium and system for transformer substation | |
JP2010276433A (en) | Imaging device, image processor, and distance measuring device | |
CN101354796A (en) | Omnidirectional stereo vision three-dimensional rebuilding method based on Taylor series model | |
Morelli et al. | Photogrammetry now and then–from hand-crafted to deep-learning tie points– | |
Guo et al. | Research on design, calibration and real-time image expansion technology of unmanned system variable-scale panoramic vision system | |
CN114092647A (en) | Three-dimensional reconstruction system and method based on panoramic binocular stereo vision | |
Wang et al. | Model-aware pre-training for radial distortion rectification | |
WO2018100230A1 (en) | Method and apparatuses for determining positions of multi-directional image capture apparatuses |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20121114 |