CN109949230A - Wide angle cameras distortion rendering method based on image recognition - Google Patents
Wide angle cameras distortion rendering method based on image recognition Download PDFInfo
- Publication number
- CN109949230A CN109949230A CN201711380839.2A CN201711380839A CN109949230A CN 109949230 A CN109949230 A CN 109949230A CN 201711380839 A CN201711380839 A CN 201711380839A CN 109949230 A CN109949230 A CN 109949230A
- Authority
- CN
- China
- Prior art keywords
- distortion
- camera
- zoom
- video camera
- graph image
- 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
Landscapes
- Image Processing (AREA)
Abstract
The wide angle cameras distortion rendering method based on image recognition that the invention discloses a kind of, this method comprises: choosing multiple zoom sampled points in the zoom variation range of camera lens；Obtain corresponding distortion of camera parameter at each zoom sampled point；By the zoom of video camera and the corresponding relationship of distortion of camera parameter, the zoom of video camera and the change curve of distortion of camera coefficient are fitted；Obtain the zoom value of current camera；According to the zoom value of the change curve and the current camera of the video camera zoom and distortion factor, current lens distortion factor is calculated；The preceding graph image size of distortion is calculated, and obtains the graph image of the size；According to the current lens distortion factor, distortion compensation is carried out to graph image virtual article in real time；Graph image virtual article after progress distortion compensation is subjected to real-time aliasing rendering with video camera real picture and is exported.
Description
Technical field
The present invention relates to computer graphic text field of image processings, and in particular to a kind of wide angle cameras based on image recognition
Distort rendering method.
Background technique
In real-time rendering technology, it will usually carry out the graph image of computer virtual and video camera acquisition signal
Aliasing generates final real time output.And during common aliasing, the virtual object view that computer generates is usually not
There is distortion effect, when such virtual object view without distortion and the scene of the video camera acquisition with distortion blend, just
Meeting is so that background mismatches before the synthesis picture generated.So that real scene and virtual pattern image on the display picture of synthesis
Correlation matching effect is very limited.
It is distorted and is not added on the image of virtual article as brought by camera lens, a true field with distortion
Scape and without distortion virtual object view aliasing when, will make aliasing export effect it is undesirable, especially when the wide-angle of camera lens
When degree is larger, effect caused by lens distortion will be more obvious, and synthesis quality will be also greatly reduced.Due to each camera shooting
The distortion of machine camera lens is a typical nonlinear optics distortion process, and the distortion effect of different camera lenses is all not quite similar, considers further that
The variation etc. of caused distortion when the zoom value variation of upper camera lens.The distortion of camera lens is to virtual scene and real scene
Bring influences either large or small, different when phase aliasing, it is difficult to the problem is handled using simple method.
Summary of the invention
For technical problem present in prior art mentioned above, the invention proposes a kind of based on image recognition
Wide angle cameras distorts rendering method, is distorted with to solve existing in the prior art as brought by camera lens and is not added
Onto the image of virtual article, a real scene with distortion and when without the virtual object view aliasing of distortion will to mix
The effect of folded output is undesirable, and especially when the wide-angle degree of camera lens is larger, effect caused by lens distortion will be brighter
It is aobvious, and synthesize the technical issues of quality will be also greatly reduced.
The wide angle cameras distortion rendering method based on image recognition of the invention includes:
In the zoom variation range of camera lens, multiple zoom sampled points are chosen；Preferably, in camera lens
In zoom variation range, at least six zoom sampled point is equally spaced chosen；
Obtain corresponding distortion of camera parameter at each zoom sampled point；
By the zoom of video camera and the corresponding relationship of distortion of camera parameter, zoom and the video camera for fitting video camera are abnormal
The change curve of variable coefficient；
Obtain the zoom value of current camera；
According to the zoom value of the change curve and the current camera of the video camera zoom and distortion factor, interpolation meter
Calculate current lens distortion factor；
The preceding graph image size of distortion is calculated, and obtains the graph image of the size；
According to the current lens distortion factor, distortion compensation is carried out to graph image virtual article in real time；
Graph image virtual article after progress distortion compensation is subjected to real-time aliasing rendering simultaneously with video camera real picture
Output.
In above-mentioned steps, first two for can choosing the coefficient of radial distortion of video camera are used as distortion of camera parameter, i.e.,
The radial distortion parameter k of radial distortion model1And k2。
Wherein, the parameter k1And k2Meet following relational expression:
Xdistorted=x (1+k1*r2+k2*r4) (1)；
Ydistorted=y (1+k1*r2+k2*r4) (2)；
Wherein x, y are coordinate of the distortionless pixel under camera coordinates system, xdistorted、ydistortedFor the picture after distortion
Coordinate of the element under camera coordinates system；k1And k2For camera lens coefficient of radial distortion, r2=x2+y2。
In alternative embodiments, the zoom sampled point and distortion parameter k are based on1And k2Corresponding relationship, using three
Secondary curve matching goes out video camera zoom and distortion factor k1And k2Change curve.
According to current distortion coefficients of camera lens k1And k2And distortion after required picture size xdistorted、ydistorted, lead to
It crosses relational expression (1) and (2) formula calculates the size x and y of required picture before distorting.
The present invention is through the above technical solutions, improving virtual engine aliasing effect makes real scene and virtual object view aliasing
When output effect, the distortion effect for really being had real scene should be added to the generation of the virtual object view of graph image
Among journey, that is to say, that when generation 3-D graphic virtual foreground signal, adequately take into account the distortion of camera lens
Effect carries out merging acquired show with camera signal by using the prospect with distortion effect same as camera signal
Show picture, simulate all processes of real camera imaging process, so that the virtual pattern image and camera shooting process that generate
It is consistent as much as possible, to greatly improve effect when not considering distortion of camera when generating computer virtual graph image
Fruit, so that the signal of aliasing synthesis, which broadcasts picture, obtains increased quality substantially.
Detailed description of the invention
Fig. 1 is the process according to a kind of wide angle cameras distortion rendering method based on image recognition of the embodiment of the present invention
Figure；
Fig. 2 is the corresponding relationship curve synoptic diagram of the zoom value and k1, k2 in the embodiment of the present invention；
Fig. 3 is picture size described in the embodiment of the present invention because of the changed schematic diagram of the effect of distortion.
Specific embodiment
To make the objectives, technical solutions, and advantages of the present invention clearer, below in conjunction with attached drawing to the present invention make into
It is described in detail to one step, the embodiments described below are only the embodiment of the present invention, are only used for clearly to this hair
It is bright to be explained and illustrated, it can not limited the scope of protection of the present invention with this.
Method of the invention is broadly divided into distortion measurement (step S101-S103) and distortion compensation (step S104-S108)
Two links, the use of the technology can allow virtual more outstanding with real integration technology.Since this method is by void
Quasi- graph image carries out distortion in real time processing, and the different camera lenses different to distortion situation can obtain preferable compensation effect.Under
Face is described for distortion measurement and distortion compensation the two links by specific example.
With reference to Fig. 1, the wide angle cameras distortion rendering method based on image recognition that the invention proposes a kind of, such as Fig. 1 institute
Show, this method comprises:
S101 chooses multiple zoom sampled points in the zoom variation range of camera lens；
S102 obtains corresponding distortion of camera parameter at each zoom sampled point；
S103, by the zoom of video camera and the corresponding relationship of distortion of camera parameter, fit the zoom value of video camera with
The change curve of distortion of camera coefficient；
S104 obtains the zoom value of current camera；
S105, according to the zoom of the change curve and the current camera of the video camera zoom value and distortion factor
Value, interpolation calculation go out current lens distortion factor；
S106 calculates graph image size before distortion, and obtains the graph image of the size；
S107 carries out distortion compensation to graph image virtual article in real time according to the current lens distortion factor；
S108, by the graph image virtual article and the real-time aliasing wash with watercolours of video camera real picture progress after progress distortion compensation
It contaminates and exports.
In the present embodiment, optionally, in step s101, in the zoom variation range of camera lens, at equal intervals
At least six zoom sampled point is chosen on ground, and sampled point should not be excessive, because sampled point can excessively make actual mechanical process become
It is comprehensive between the convenience and accuracy of estimation distortion Long-term change trend in precision improvement that is cumbersome, and will not bringing too many
It accepts or rejects, 6 sampled points can be chosen to describe the nonlinear distortion overall trend.It is abnormal when usual zoom lens zoom variation
Becoming characteristic also can and then change.The variation tendency that distorts when in order to obtain zoom variation, zoom can be divided for several segments, according to
Distortion situation in each section of zoom carries out interpolation to distortion to continuously acquire the distortion situation under each zoom.
In step s 102, for each zoom sampled point, the acquisition of Zhang Zhengyou camera calibration method specifically can be used
Corresponding distortion of camera parameter at the zoom sampled point.Because lens distortion is nonlinear distortion, usually that this is non-linear
Distortion is with Taylor series expansion come approximate description.Due to the presence of error etc., if the coefficient selects excessively bring more preferably
As a result, and calculation amount increase instead it is very more.In general it can assume that lens distortion is symmetrically become around along optical center
Change, so Taylor expansion item only has even order terms.First two in the coefficient of radial distortion of this selection camera are used as distortion ginseng
Number, the i.e. k of radial distortion model1、k2Two parameters, shown in following formula.
xdistorted=x (1+k1*r2+k2*r4)(1)；
ydistorted=y (1+k1*r2+k2*r4)(2)；
Wherein x, y are coordinate of the distortionless pixel under camera coordinates system, xdistorted、ydistortedFor the picture after distortion
Coordinate of the element under camera coordinates system；k1And k2For camera lens coefficient of radial distortion, r2=x2+y2。
Wherein in step s 103, at least 6 groups of zoom and distortion parameter k acquired in step s1011、k2Corresponding close
System, optional Cubic Curve Fitting go out video camera zoom and distortion factor k1、k2Change curve, i.e. k1=f1(zoom)、k2=
f2(zoom), as shown in Figure 2.
Step S101-S103 is mainly the method by camera calibration, obtains zoom and distortion factor k1、k2Relationship
Curve is used for obtaining distortion factor corresponding to any zoom value.Obtaining video camera zoom and distortion factor k1、k2
Relation curve f1、f2Later, that is, the zoom value interpolation of current camera can be used to obtain distortion factor, and in real time to figure
Image virtual object view carries out distortion compensation, so that virtual object view signal and video camera real scene signal obtain preferably when blending
Aliasing effect, specially step S104-S108.
In step s105, after getting the zoom value of current camera, the zoom value is brought into function f1、f2In, meter
Calculate current distortion coefficients of camera lens k1、k2。
In step s 106, according to distortion coefficients of camera lens k at this time1、k2And distortion after required graph image size
xdistorted、ydistorted, size x, y of graph image before distorting needed for being gone out by formula (1) and (2) inverse.And according to
This result obtains the graph image of the size.This is the figure in order to guarantee distortion treated graph image and camera signal
The length and width pixel of image is consistent, because the effect meeting of distortion is so that the size of graph image changes, as shown in Figure 3.
In step s 107, according to distortion factor at this time, optionally using bilinear interpolation method to virtual object view figure
As applying distortion effect, and will distortion treated that graph image is converted into and the consistent graphic diagram of camera signal picture size
Picture, in order to synthesize it with the live signal of video camera shooting.
Finally, in step S108, graphical virtual picture signal and video camera actual signal are subjected to real-time aliasing and defeated
Out.
A specific embodiment of the invention is described in detail above, but those skilled in the art are according to this
The creative concept of invention can carry out various changes and modifications to the present invention, but the various changes and modifications done do not depart from
The spirit and scope of the present invention, within the scope of coming under the claims in the present invention.
Claims (6)
- The rendering method 1. a kind of wide angle cameras based on image recognition distorts, which is characterized in that this method comprises:In the zoom variation range of camera lens, multiple zoom sampled points are chosen；Obtain corresponding distortion of camera parameter at each zoom sampled point；By the zoom of video camera and the corresponding relationship of distortion of camera parameter, zoom and the distortion of camera system of video camera are fitted Several change curves；Obtain the zoom value of current camera；According to the zoom value of the change curve and the current camera of the video camera zoom value and distortion factor, calculates and work as Preceding distortion coefficients of camera lens；The preceding graph image size of distortion is calculated, and obtains the graph image of the size；According to the current lens distortion factor, distortion compensation is carried out to graph image in real time；Graph image virtual article after progress distortion compensation is subjected to real-time aliasing rendering with video camera real picture and is exported.
- 2. the method according to claim 1, wherein in the zoom variation range of camera lens, at equal intervals Choose at least six zoom sampled point in ground.
- 3. the method according to claim 1, wherein first two that choose the coefficient of radial distortion of camera are used as and take the photograph Camera distortion parameter, i.e. the radial distortion parameter k of radial distortion model1And k2。
- 4. according to the method described in claim 3, it is characterized in that, the parameter k1And k2Meet following relational expression:Xdistorted=x (1+k1*r2+k2*r4) (1)；Ydistorted=y (1+k1*r2+k2*r4) (2)；Wherein x, y are coordinate of the distortionless pixel under camera coordinates system, xdistorted、ydistortedExist for the pixel after distortion Coordinate under camera coordinates system；k1And k2For camera lens coefficient of radial distortion, r2=x2+y2。
- 5. according to the method described in claim 4, it is characterized in that, being based on the zoom sampled point and distortion parameter k1And k2's Corresponding relationship goes out video camera zoom and distortion factor k using Cubic Curve Fitting1And k2Change curve.
- 6. according to the method described in claim 4, it is characterized in that, according to current distortion coefficients of camera lens k1And k2And distortion The size x of required picture afterwardsdistorted、ydistorted, picture is big before distorting needed for being calculated by relational expression (1) and (2) formula Small x and y.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711380839.2A CN109949230A (en) | 2017-12-20 | 2017-12-20 | Wide angle cameras distortion rendering method based on image recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711380839.2A CN109949230A (en) | 2017-12-20 | 2017-12-20 | Wide angle cameras distortion rendering method based on image recognition |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109949230A true CN109949230A (en) | 2019-06-28 |
Family
ID=67005554
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711380839.2A Pending CN109949230A (en) | 2017-12-20 | 2017-12-20 | Wide angle cameras distortion rendering method based on image recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109949230A (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101324433A (en) * | 2007-06-14 | 2008-12-17 | 佳能株式会社 | Method and apparatus for measuring position and orientation of an object |
| CN101520897A (en) * | 2009-02-27 | 2009-09-02 | 北京机械工业学院 | Video camera calibration method |
| CN101814186A (en) * | 2010-02-04 | 2010-08-25 | 上海交通大学 | Method utilizing curve-fitting to calibrate radial distortion of camera |
| CN102075785A (en) * | 2010-12-28 | 2011-05-25 | 武汉大学 | Method for correcting wide-angle camera lens distortion of automatic teller machine (ATM) |
| CN102857701A (en) * | 2012-09-14 | 2013-01-02 | 北京东方艾迪普科技发展有限公司 | Method for tracking virtual camera in three-dimensional scene |
| CN104657940A (en) * | 2013-11-22 | 2015-05-27 | 中兴通讯股份有限公司 | Method and device for correction remediation and analysis alarm of distorted image |
| US20150365605A1 (en) * | 2009-03-19 | 2015-12-17 | Digitaloptics Corporation | Dual sensor camera |
| CN106408542A (en) * | 2016-10-10 | 2017-02-15 | 四川大学 | Rapid geometric correction method of dome visual scene |
-
2017
- 2017-12-20 CN CN201711380839.2A patent/CN109949230A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101324433A (en) * | 2007-06-14 | 2008-12-17 | 佳能株式会社 | Method and apparatus for measuring position and orientation of an object |
| CN101520897A (en) * | 2009-02-27 | 2009-09-02 | 北京机械工业学院 | Video camera calibration method |
| US20150365605A1 (en) * | 2009-03-19 | 2015-12-17 | Digitaloptics Corporation | Dual sensor camera |
| CN101814186A (en) * | 2010-02-04 | 2010-08-25 | 上海交通大学 | Method utilizing curve-fitting to calibrate radial distortion of camera |
| CN102075785A (en) * | 2010-12-28 | 2011-05-25 | 武汉大学 | Method for correcting wide-angle camera lens distortion of automatic teller machine (ATM) |
| CN102857701A (en) * | 2012-09-14 | 2013-01-02 | 北京东方艾迪普科技发展有限公司 | Method for tracking virtual camera in three-dimensional scene |
| CN104657940A (en) * | 2013-11-22 | 2015-05-27 | 中兴通讯股份有限公司 | Method and device for correction remediation and analysis alarm of distorted image |
| CN106408542A (en) * | 2016-10-10 | 2017-02-15 | 四川大学 | Rapid geometric correction method of dome visual scene |
Non-Patent Citations (4)
| Title |
|---|
| 165067718: "曲线拟合在任意焦距的畸变校正中的应用", 《道客巴巴》 * |
| 张佳成等: "图像几何畸变校正方法", 《计算机工程与应用》 * |
| 张涛: "基于光学系统畸变设计曲线的畸变校正方法", 《智能计算机与应用》 * |
| 高宏伟等: "《电子封装工艺与装备技术基础教程》", 30 June 2017, 《西安电子科技大学出版社》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Chen et al. | Camera lens super-resolution | |
| US20050007477A1 (en) | Correction of optical distortion by image processing | |
| CN104537616A (en) | Correction Method of Fisheye Image Distortion | |
| JP5696783B2 (en) | Image processing device | |
| JP2013211827A (en) | Image processing method, device and program | |
| CN102930507A (en) | Image processing method, image processing apparatus, and image pickup apparatus | |
| CN110149511B (en) | Distortion correction method, distortion correction device and display system | |
| CN105472265B (en) | A kind of device and method obtaining high dynamic range images | |
| CN104182969B (en) | Internal and external parameter calibration method of single-scanline camera | |
| WO2023160426A1 (en) | Video frame interpolation method and apparatus, training method and apparatus, and electronic device | |
| CN107680140A (en) | A kind of depth image high-resolution reconstruction method based on Kinect cameras | |
| CN111932453B (en) | High-resolution image generation method and high-speed camera integrating same | |
| JP6071860B2 (en) | Image processing method, image processing apparatus, imaging apparatus, and image processing program | |
| CN106468562B (en) | A kind of color camera radial direction color difference calibration method based on absolute phase | |
| CN104735351A (en) | High resolution light field image recreation method and imaging device | |
| CN106817575A (en) | The method of the method and automatic corrigendum of image capturing device and its generation depth information | |
| CN106920216A (en) | A kind of method and device for eliminating picture noise | |
| CN112465705B (en) | Visual field expanding system and method based on two-aperture rotating biprism | |
| CN109949230A (en) | Wide angle cameras distortion rendering method based on image recognition | |
| CN204287060U (en) | A kind of online three-dimensional detection device based on RGB structured light | |
| Petrov et al. | Color correction with 3D lookup tables in diffractive optical imaging systems | |
| JP2003169255A (en) | Shading correction method | |
| Liu et al. | A time-domain projection-based registration-scene-based nonuniformity correction technology and its detailed hardware realization | |
| CN116503248A (en) | Infrared image correction method and system for crude oil storage tank | |
| JP2017103695A (en) | Image processing apparatus, image processing method, and program of them |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: 100101 Beijing city Chaoyang District District three building 10, building 1, Hui Li two Applicant after: Aidipu Technology Co., Ltd Address before: 100101 Beijing city Chaoyang District District three building 10, building 1, Hui Li two Applicant before: IDEAPOOL (BEIJING) CULTURE AND TECHNOLOGY Co.,Ltd. |
|
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190628 |