CN102496178A - Three-dimensional smoke density field generating method based on single-viewpoint images - Google Patents
Three-dimensional smoke density field generating method based on single-viewpoint images Download PDFInfo
- Publication number
- CN102496178A CN102496178A CN2011103285466A CN201110328546A CN102496178A CN 102496178 A CN102496178 A CN 102496178A CN 2011103285466 A CN2011103285466 A CN 2011103285466A CN 201110328546 A CN201110328546 A CN 201110328546A CN 102496178 A CN102496178 A CN 102496178A
- Authority
- CN
- China
- Prior art keywords
- smog
- density field
- smoke
- dimensional
- dimensional density
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Landscapes
- Processing Or Creating Images (AREA)
Abstract
The invention relates to a three-dimensional smoke density field generating method based on single-viewpoint images. Three-dimensional smoke density field can be regenerated on a computer by the aid of data of smoke images shot from a single viewpoint to create three-dimensional animation of smoke. The method includes: firstly, using the image segmentation method to segment existing smoke data into local smokes; secondly, computing principal direction and outline basis function of each local smoke according to the local smokes; thirdly, using the proceduring method to generate the three-dimensional density fields of the local smokes according to the principal direction and outline basis function of each local smoke; and fourthly, joining the three-dimensional density fields to generate a continuous three-dimensional smoke density field. The three-dimensional smoke density field generating method based on single-viewpoint images has the advantages that firstly, data acquisition is simple, and only one camera is used for acquiring smoke data from one viewpoint; and secondly, processing is quick, and the three-dimensional smoke density field can be generated in real time.
Description
Technical field
The invention belongs to the computer animation field, specifically, is a kind of method that generates the three-dimensional density field of smog based on the single view image.
Background technology
The modeling of sense of reality object is the emphasis of graphics research with drawing always, also is key issue in application contents such as virtual reality.In recent years, the active research focus that becomes computer graphics and computer vision field based on the real-world object modeling and the method for reconstructing of camera.Modeling based on image has the higher sense of reality, especially for the complex natural phenomenon, like motion of reflecting attribute, people and the clothing of the geometric configuration of trees and flower, people's hair and skin etc.These phenomenons or object often are difficult to efficiently really come modeling through traditional modeling pattern such as analytic expression, mathematical models owing to its complicacy.
Participate in medium, like cigarette, mist, atmosphere etc., the sense of reality of enhanced scene fully in virtual environment.Light is propagated in participating in medium, has experienced the illumination variation of rich and complex, gives the strong sense of reality and feeling of immersion.At present, the geometrical Modeling Technology to solid body has obtained a lot of achievements in research; Compare for the modeling work of participation medium with dynamic fluid characteristic such as smog that the former is then less.The vision of smog is obtained theoretical method and the gordian technique with modeling in the research real world, and the application that promotes sense of reality attribute to be modeled in the digitizing technique field is had important scientific meaning and using value.
The modeling of participating in medium is the focus and the difficult point of field of Computer Graphics in recent years with drawing.For participation medium heterogeneous, like the cigarette of motion, its modeling pattern mainly is based on physical simulation for a long time.According to original state and physical equation, draw the density field data of cigarette on each timing node with the method for numerical evaluation.Its shortcoming also is very tangible, and the numerical calculations amount is big on the one hand, and each frame need be separated PDE, and very sensitive for parameter; On the other hand, for the minutia of complicacy of the cigarette of motion, what the method for simulation can't be detailed shows.Based on the method for image the processing of these aspects had inborn advantage; Directly remove to gather and catch real medium; Abundant ins and outs in reservation real phenomena that can be complete or the real-world object, and calculating strength also less than or be equivalent to method based on the numerical simulation of physics.
Summary of the invention
The purpose of this invention is to provide the generation method of the three-dimensional density field of a kind of smog based on the single view image, use the smog data of gathering,, make up the three-dimensional density field of smog, realize modeling smog through procedural method from single view.
Another purpose of this method provides a kind of smog modeling pattern easily and fast, has realized the smog in the true environment is carried out rapid modeling.
For achieving the above object, the present invention proposes a kind of generation method of the three-dimensional density field of smog based on the single view image, and concrete way is following:
1) method to existing smog The data image segmentation is divided into local smoke;
2) be directed against principal direction and the profile basis function that each local smoke is calculated this part smog;
3) utilize the three-dimensional density field of procedural method according to the principal direction and the profile basis function generation local smoke of each local smoke;
4) the three-dimensional density field that generates is seamlessly transitted, generate the three-dimensional density field of continuous smog.
In the technique scheme, in the step 1), existing smog data collect from single viewpoint direction through the single image collecting device.
In the technique scheme, step 2) in, comprise to local smoke and calculate the principal direction of this part smog and the method for profile basis function.
In the technique scheme, in the step 3), generate the procedural method of the three-dimensional density field of local smoke according to the principal direction of each local smoke and profile basis function.
In the technique scheme, in the step 4), described transition comprises carries out seamlessly transitting of time and space respectively to the smog data that have transformational relation.
The invention has the advantages that:
1, data acquisition is simple, only gathers the smog data with single camera from a viewpoint and gets final product;
2, the processing time quick, can generate the three-dimensional density field of smog in real time.
Description of drawings
Fig. 1 illustrates implementing procedure figure of the present invention.
Embodiment
Below in conjunction with accompanying drawing and embodiment the present invention is done further explain.
Main process flow diagram of the present invention is as shown in Figure 1, and concrete steps are following:
(1) method to existing smog The data image segmentation is divided into local smoke.
Wherein, existing smog data collect from single viewpoint direction through the single image collecting device.The image capture device here can be digital camera, video camera or high speed camera.
The image partition method step comprises three steps:
1.1; Calculate binary map by the smog view data that collects; And calculate distance map through range conversion according to binary map; Method be by boundary to inside assignment successively, the distance value of treating the assignment pixel adds 1 for known maximum distance value in its neighbor pixel, puts the distance on smog border in order to expression smog interior pixels;
1.2, obtaining the cluster point of local smoke by distance map, method is to find the point that has local maximum in the distance map, there is a middle distance smog border in these institutes of putting in the close region around farthest, and these points are appointed as the cluster point;
1.3, utilize the cluster smog image of naming a person for a particular job to cut apart.
In step (1),, in the gatherer process of data, can use high speed camera to carry out data acquisition as far as possible in order to obtain the three-dimensional density field of desirable continuous smog.Gathering environment is color background as far as possible, and the purpose of doing like this is the operation that makes things convenient for image segmentation step, guarantees to obtain ideal results.
(2) be directed against principal direction and the profile basis function that each local smoke is calculated this part smog.
In this step, the method for principal direction and profile basis function of calculating this part smog to each local smoke is following:
At first, local smoke image pixel coordinate is sampled, sampled data is carried out principal component analysis (PCA).The principal direction that obtains is the principal direction of this part smog of asking.The local smoke image space is transformed in the principal direction coordinate system, and in the new coordinate system after conversion, the principal direction of smog image drops on the x axle.In new coordinate system on/pixel of lower boundary samples, and select and have local maximum/minimizing pixel.Sampled result is carried out the function match, promptly obtain profile basis function f
+And f
-, the profile basis function representes that on the principal direction each put the ultimate range on border.
(3) utilize the three-dimensional density field of procedural method according to the principal direction and the profile basis function generation local smoke of each local smoke.
In this step, it is following to utilize procedural method to generate the method for three-dimensional density field of local smoke according to the principal direction of each local smoke and profile basis function:
At first, (x y), calculates y coordinate smallest point (x, y with identical x coordinate figure to the arbitrfary point p in the local smoke image
m) and maximum point (x, y
M).
Secondly, by the profile basis function f that obtains in the step (2)
+And f
-, (x y) calculates f to a p
+(x) and f
-(x), obtain the distance of principal direction to up-and-down boundary:
To a p (x, y) interpolation (y
m, f
-And (y (x))
M, f
+(x))
l＝(1-α)f
-(x)+αf
+(x)
Obtain a p to principal direction apart from l and and then the depth information d of calculation level p.
Make up the three-dimensional density field of this local smoke by the depth information of being had a few in this local smoke image:
Depth information d representes the distance of a p to the image imaging plane, and the XY plane of three-dimensional density field is appointed as on the image imaging plane, can be obtained the three-dimensional coordinate of a p by former images acquired two-dimensional coordinate and depth information d.The density value of representing this smog with the pixel value of former images acquired corresponding point.Obtain the three-dimensional density field of local smoke thus.
(4) the three-dimensional density field that generates is seamlessly transitted, generate the three-dimensional density field of continuous smog.
In this step, according to the three-dimensional density field of the smog that generates in the step (3), we accomplish the modeling to single frames smog.Further, we also need carry out seamlessly transitting of time and space respectively to the three-dimensional density field of the smog between frame and the frame, have so just intactly created continuous smoke density field.
The method that the three-dimensional density field that generates is seamlessly transitted is following:
At first, confirm position in the time series of the frame of asking between the key frame of front and back.
Respectively by the principal direction of front and back key frame and the principal direction and the profile basis function of profile basis function interpolation calculation present frame.Principal direction is by vector representation, and the principal direction of the frame of asking can be obtained by the principal direction vector interpolation of front and back key frame.The profile basis function is confirmed by the key point of choosing in the border; Each group key point can be confirmed a profile basis function; Earlier go out the key point in the frame of asking, carry out the function match by these key points again and obtain the profile basis function by the key point interpolation calculation in the key frame border, front and back.By principal direction that obtains and profile basis function, utilize the method for step (3), generate the three-dimensional density field of smog of present frame.
Claims (5)
1. generation method based on the three-dimensional density field of the smog of single view image is characterized in that may further comprise the steps:
1) method to existing smog The data image segmentation is divided into local smoke;
2) be directed against principal direction and the profile basis function that each local smoke is calculated this part smog;
3) utilize the three-dimensional density field of procedural method according to the principal direction and the profile basis function generation local smoke of each local smoke;
4) the three-dimensional density field that generates is seamlessly transitted, generate the three-dimensional density field of continuous smog.
2. the generation method of the three-dimensional density field of the smog based on the single view image according to claim 1, it is characterized in that: existing smog data collect from single viewpoint direction through the single image collecting device.
3. according to the generation method of the three-dimensional density field of the described smog based on the single view image of claim 1; It is characterized in that: step 2) in, cut apart at first the smog image being carried out figure to the principal direction of this part smog of local smoke calculating and the method for profile basis function; Principal component analysis (PCA) is carried out in sampling to local smoke then, calculates principal direction; And according to principal direction generation profile basis function.
4. according to the generation method of the three-dimensional density field of the described smog based on the single view image of claim 1; It is characterized in that: in the step 3); The procedural method that generates the three-dimensional density field of local smoke according to the principal direction of each local smoke and profile basis function is according to the smog image; Calculate the distance that this puts principal direction to being judged to be the inner point of smog by the profile basis function, calculate the depth information of this point then, obtain three-dimensional density field.
5. according to the generation method of the three-dimensional density field of the described smog based on the single view image of claim 1, it is characterized in that: seamlessly transitting described in the step 4) comprises carries out seamlessly transitting of time and space respectively to the smog data that have transformational relation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110328546.6A CN102496178B (en) | 2011-10-26 | 2011-10-26 | Three-dimensional smoke density field generating method based on single-viewpoint images |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110328546.6A CN102496178B (en) | 2011-10-26 | 2011-10-26 | Three-dimensional smoke density field generating method based on single-viewpoint images |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102496178A true CN102496178A (en) | 2012-06-13 |
| CN102496178B CN102496178B (en) | 2014-06-25 |
Family
ID=46188000
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201110328546.6A Expired - Fee Related CN102496178B (en) | 2011-10-26 | 2011-10-26 | Three-dimensional smoke density field generating method based on single-viewpoint images |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102496178B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103886636A (en) * | 2014-01-28 | 2014-06-25 | 浙江大学 | Real-time smoke rendering algorithm based on ray cast stepping compensation |
| CN104156995A (en) * | 2014-07-16 | 2014-11-19 | 浙江大学 | Production method for ribbon animation aiming at Dunhuang flying image |
| CN104361629A (en) * | 2014-12-01 | 2015-02-18 | 北京航空航天大学 | Smoke model space editing method based on streamline deformation |
| CN110084872A (en) * | 2019-03-25 | 2019-08-02 | 中国科学院计算技术研究所 | A kind of the Animation of Smoke synthetic method and system of data-driven |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH10132807A (en) * | 1996-10-29 | 1998-05-22 | Mitsubishi Rayon Co Ltd | Evaluation method for smoke of cigarette |
| CN1404019A (en) * | 2002-10-23 | 2003-03-19 | 北京航空航天大学 | Method of creating vivid lighting effect under virtual environment several factor effects |
| US20090006047A1 (en) * | 2007-06-26 | 2009-01-01 | Microsoft Corporation | Real-Time Rendering of Light-Scattering Media |
-
2011
- 2011-10-26 CN CN201110328546.6A patent/CN102496178B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH10132807A (en) * | 1996-10-29 | 1998-05-22 | Mitsubishi Rayon Co Ltd | Evaluation method for smoke of cigarette |
| CN1404019A (en) * | 2002-10-23 | 2003-03-19 | 北京航空航天大学 | Method of creating vivid lighting effect under virtual environment several factor effects |
| US20090006047A1 (en) * | 2007-06-26 | 2009-01-01 | Microsoft Corporation | Real-Time Rendering of Light-Scattering Media |
Non-Patent Citations (3)
| Title |
|---|
| JIAPING WANG等: "Modeling and Rendering of Heterogeneous Translucent Materials Using the Diffusion Equation", 《ACM TRANSACTIONS ON GRAPHICS》 * |
| MOHIT GUPTA等: "Legendre Fluids: A Unified Framework for Analytic Reduced Space Modeling and Rendering of Participating Media", 《EUROGRAPHICS/ACM SIGGRAPH SYMPOSIUM ON COMPUTER ANIMATION (2007)》 * |
| 胡勇 等: "一种基于图像的非均匀单散射参与介质建模方法", 《中国科学：信息科学》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103886636A (en) * | 2014-01-28 | 2014-06-25 | 浙江大学 | Real-time smoke rendering algorithm based on ray cast stepping compensation |
| CN104156995A (en) * | 2014-07-16 | 2014-11-19 | 浙江大学 | Production method for ribbon animation aiming at Dunhuang flying image |
| CN104361629A (en) * | 2014-12-01 | 2015-02-18 | 北京航空航天大学 | Smoke model space editing method based on streamline deformation |
| CN104361629B (en) * | 2014-12-01 | 2017-03-29 | 北京航空航天大学 | A kind of cigarette model space edit methods deformed based on streamline |
| CN110084872A (en) * | 2019-03-25 | 2019-08-02 | 中国科学院计算技术研究所 | A kind of the Animation of Smoke synthetic method and system of data-driven |
| CN110084872B (en) * | 2019-03-25 | 2020-12-25 | 中国科学院计算技术研究所 | Data-driven smoke animation synthesis method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102496178B (en) | 2014-06-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109191548A (en) | Animation method, device, equipment and storage medium | |
| CN103473801B (en) | A kind of human face expression edit methods based on single camera Yu movement capturing data | |
| CN104915978B (en) | Realistic animation generation method based on body-sensing camera Kinect | |
| CN109147017A (en) | Dynamic image generation method, device, equipment and storage medium | |
| Xie et al. | Tree modeling with real tree-parts examples | |
| CN111043988B (en) | Single stripe projection measurement method based on graphics and deep learning | |
| CN102496178B (en) | Three-dimensional smoke density field generating method based on single-viewpoint images | |
| CN105427386A (en) | Garment deformation method based on input human body posture real-time generation | |
| CN107613161A (en) | Video data handling procedure and device, computing device based on virtual world | |
| CN110751730A (en) | Dressing human body shape estimation method based on deep neural network | |
| CN106023316A (en) | Kinect-based dynamic sequence capture method | |
| CN110503707A (en) | A kind of true man's motion capture real-time animation system and method | |
| CN110298917B (en) | Face reconstruction method and system | |
| CN105719279B (en) | Based on the modeling of cylindroid trunk and arm regions segmentation and arm framework extraction method | |
| CN114359445A (en) | Training data generation method and device for human body bone joint point extraction model | |
| CN107633547A (en) | Realize the view data real-time processing method and device, computing device of scene rendering | |
| Xu et al. | 3D joints estimation of the human body in single-frame point cloud | |
| CN107644686A (en) | Medical data acquisition system and method based on virtual reality | |
| CN106504063B (en) | A kind of virtual hair tries video frequency showing system on | |
| He et al. | Sketch2cloth: Sketch-based 3d garment generation with unsigned distance fields | |
| CN103886293B (en) | Human body behavior recognition method based on history motion graph and R transformation | |
| CN113763536A (en) | Three-dimensional reconstruction method based on RGB image | |
| CN113920020A (en) | Human point cloud real-time repairing method based on depth generation model | |
| Kuang et al. | A Pixel image generation algorithm based on CycleGAN | |
| Jiang et al. | Animation scene generation based on deep learning of CAD data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20140625 Termination date: 20171026 |