CN101833786B - Method and system for capturing and rebuilding three-dimensional model - Google Patents
Method and system for capturing and rebuilding three-dimensional model Download PDFInfo
- Publication number
- CN101833786B CN101833786B CN2010101411826A CN201010141182A CN101833786B CN 101833786 B CN101833786 B CN 101833786B CN 2010101411826 A CN2010101411826 A CN 2010101411826A CN 201010141182 A CN201010141182 A CN 201010141182A CN 101833786 B CN101833786 B CN 101833786B
- Authority
- CN
- China
- Prior art keywords
- dimensional model
- static
- model
- visual
- constraint
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 62
- 230000003068 static effect Effects 0.000 claims abstract description 60
- 230000000007 visual effect Effects 0.000 claims abstract description 51
- 230000033001 locomotion Effects 0.000 claims description 40
- 239000011159 matrix material Substances 0.000 claims description 16
- 230000003287 optical effect Effects 0.000 claims description 8
- 238000005457 optimization Methods 0.000 claims description 3
- 238000011084 recovery Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 2
- 238000002156 mixing Methods 0.000 abstract 1
- 238000013461 design Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 230000007774 longterm Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Landscapes
- Processing Or Creating Images (AREA)
- Image Processing (AREA)
Abstract
The invention provides a method for capturing and rebuilding a static three-dimensional model, which comprises the following steps of: acquiring an image of a moving object in an annular field; acquiring a visible shell model; acquiring a depth point cloud of each visual angle according to each visual image, the visible shell model and a preset constraint condition; and blending the depth point cloud of each visual angle to obtain the static three-dimensional model. The method of the invention can guarantee the precision and integrity of the rebuilt shape of the static three-dimensional model. In addition, the invention also provides a method for capturing and rebuilding a dynamic three-dimensional model.
Description
Technical Field
The invention relates to the technical field of computer video processing, in particular to a method and a system for capturing and reconstructing a three-dimensional model.
Background
For the three-dimensional reconstruction problem of a dynamic scene, a lot of work regards the problem as simple accumulation of the static scene reconstruction problem on a time dimension, namely, the scene reconstruction is not assisted by time information, and each frame is independently subjected to static three-dimensional modeling. However, the method has high complexity and large storage capacity, cannot ensure the topological consistency of the models between frames, and is easy to generate a jitter phenomenon. In addition, the three-dimensional modeling by adopting the method cannot effectively analyze the motion condition of the non-rigid model, and cannot obtain a model at any moment through interpolation in a time domain. By studying such problems, the prior art proposes reconstruction methods that jointly solve 3D scene streams and geometric models. Furthermore, the method for reconstructing the geometry and the motion of the dynamic scene by using the variation method is also provided, however, the geometric reconstruction and the motion reconstruction are performed iteratively, that is, the geometric reconstruction at a certain moment is used as an initial value of the motion reconstruction to derive the model reconstruction at the next moment, so that the space-time joint reconstruction efficiency of the method is still not high, and the actual effect is not satisfactory.
Therefore, in order to avoid the problems of high difficulty and general quality of space-time joint reconstruction, another type of video-based dynamic three-dimensional reconstruction method takes a static three-dimensional reconstruction result of an initial frame as a scene representation, then uses a three-dimensional motion tracking algorithm to solve the motion of the three-dimensional object, and uses a proper deformation algorithm to drive a static model to move, thereby obtaining a dynamic three-dimensional reconstruction result. Currently, video-based three-dimensional motion tracking can be divided into two categories: tagged three-dimensional motion tracking and untagged three-dimensional motion tracking. Among them, the marked three-dimensional motion tracking method is accurate, but requires the close-fitting garment with the mark to be worn by the capturing actor, thereby limiting the capture of shape and texture. The three-dimensional motion tracking method without the mark overcomes the defects. One label-free three-dimensional motion tracking method captures the motion of a human body wearing more general clothing by combining a kinematic model and a clothing model, but this method cannot capture the precise geometry of a moving object. Another label-free three-dimensional motion tracking method can capture the motion of the skeleton and shape of the object at the same time, but the method still cannot effectively perform three-dimensional motion tracking because some local surfaces do not change due to time. Furthermore, since the method relies only on contour information, it is very sensitive to contour errors. Although this unmarked method has increased flexibility, it is difficult to achieve the same accuracy as the marked method. In addition, most three-dimensional motion tracking methods need to help capture motion by extracting a kinematic skeleton that can only track rigid motion, so such methods often require other scanning techniques to assist in capturing time-varying shapes. Finally, all of the above methods fail to track the movement of a person wearing any apparel.
In recent years, new methods of motion picture capture and design, motion picture editing, and transformation transmission have been emerging in computer graphics. These methods no longer rely on kinematic skeletons and kinematic parameters, but are based on surface models and general shape deformation methods, so that both rigid and non-rigid deformations can be captured. However, in all such motion capture and recovery methods based on multi-view video, the static three-dimensional reconstruction of the initial frame needs to be performed by using a laser scanner. Although laser scanners can achieve high accuracy three-dimensional reconstruction results, laser scanners are expensive, time consuming and labor intensive, and the person must be completely stationary during scanning. Moreover, for the convenience of subsequent work, a person usually stands with both hands holding a fist, and the shot multi-view video is also used for doing actions with both hands holding a fist. In addition, by using the reconstruction result of the laser scanner as the initial scene representation, some surface features on the model during scanning, such as folds of clothes, and the like, are kept in the recovered whole dynamic three-dimensional sequence.
Disclosure of Invention
The present invention aims to solve at least the above technical drawbacks and proposes a method and a system for capturing and reconstructing static and dynamic three-dimensional models.
In order to achieve the above object, the present invention provides a method for capturing and reconstructing a static three-dimensional model, comprising the following steps: carrying out image acquisition on a moving object in the annular field; acquiring a visual shell model; obtaining depth point clouds of all the visual angles according to the images of all the visual angles, the visual shell model and preset constraint conditions; and fusing the obtained depth point clouds of all the visual angles to obtain a static three-dimensional model.
The invention also provides a system for capturing and reconstructing the static three-dimensional model, which comprises: the cameras surround the annular field and are used for acquiring images of a moving object in the annular field; the static three-dimensional model reconstruction device is used for acquiring a visual shell model, acquiring depth point clouds of all visual angles according to images of all visual angles, the visual shell model and preset constraint conditions, and fusing the acquired depth point clouds of all visual angles to obtain a static three-dimensional model.
The invention also provides a method for capturing and reconstructing the dynamic three-dimensional model, which comprises the following steps: obtaining a static three-dimensional model; converting a surface model of the static three-dimensional model into a body model, and using the body model as a default scene representation of motion tracking; acquiring initial three-dimensional motion of a model vertex at the next moment; selecting an accurate vertex from the obtained vertexes as position constraint of body deformation according to a preset space-time constraint condition; and updating the dynamic three-dimensional model according to the position constraint driving Laplace body deformation framework.
In another aspect, the present invention further provides a system for capturing and reconstructing a dynamic three-dimensional model, including: the cameras surround the annular field and are used for acquiring images of a moving object in the annular field; the static three-dimensional model acquisition device is used for acquiring a static three-dimensional model; and the dynamic three-dimensional model reconstruction device is used for converting the surface model of the static three-dimensional model into a body model, representing the body model as a default scene of motion tracking, acquiring initial three-dimensional motion of model vertexes at the next moment, selecting an accurate vertex from the acquired vertexes as position constraint of body deformation according to a preset space-time constraint condition, and driving the Laplace body deformation framework to update the dynamic three-dimensional model according to the position constraint.
The invention can ensure the accuracy and the integrity of the reconstructed shape of the static three-dimensional model, and in addition, the invention designs a new three-dimensional motion estimation method based on a sparse representation theory and a deformation optimization frame based on a body model, thereby obtaining a high-quality dynamic reconstruction result. In addition, the invention can be independent of three-dimensional scanners and optical markers, thus having low cost and being capable of tracking the movement of people wearing any dress.
Additional aspects and advantages of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
The foregoing and/or additional aspects and advantages of the present invention will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
FIG. 1 is a flowchart of a method for capturing and reconstructing a static three-dimensional model according to an embodiment of the present invention;
FIG. 2 shows 20 cameras annularly distributed around a scene to be captured according to an embodiment of the present invention;
FIG. 3 is a flowchart of a method for capturing and reconstructing a dynamic three-dimensional model according to an embodiment of the present invention;
FIG. 4 is a schematic block diagram of the entire dynamic three-dimensional reconstruction method according to an embodiment of the present invention; and
FIG. 5 shows the results of a dynamic three-dimensional model obtained by applying the method of the present invention to two long-term sequences.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are illustrative only and should not be construed as limiting the invention.
The embodiments of the present invention respectively provide methods for capturing and reconstructing a static three-dimensional model and a dynamic three-dimensional model, but it should be noted that the capturing and reconstructing of a dynamic three-dimensional model may be based on a static three-dimensional model obtained by the present invention, or may be based on a static three-dimensional model obtained by other means, such as an existing three-dimensional scanner, and these methods are all included in the protection scope of the present invention.
As shown in fig. 1, a flowchart of a method for capturing and reconstructing a static three-dimensional model according to an embodiment of the present invention includes the following steps:
and step S101, carrying out image acquisition on the moving object in the annular field. For example, 20 cameras are arranged in the annular field, the frame rate of each camera is 30 frames/second, and the cameras in each group are controlled to acquire moving objects in the annular field. Of course, the skilled person can select more cameras to obtain more view images, and of course, the number of cameras can be reduced, which are all included in the scope of the present invention. An example of the present invention, as shown in fig. 2, 20 cameras are distributed annularly around a scene to be captured according to an embodiment of the present invention. Where Ci denotes camera No. i. The resolution of the images acquired by the camera is 1024 × 768. The collected character stands at the center of the ring.
In step S102, a visual shell model (visual hull) at the initial time is acquired.
And S103, obtaining depth point clouds of all the visual angles according to the images of all the visual angles, the visual shell model and a preset constraint condition. The method specifically comprises the following steps:
step S201, intersecting the image of each viewing angle with the obtained visual shell model to obtain a visible point cloud of each viewing angle.
In step S202, the visible point cloud of each view angle is projected to the view angle image, and an initial depth point cloud estimate is obtained, where d is (a, b, 1) as a deviation along the epipolar line direction.
Step S203, obtaining an accurate depth point cloud according to the initial depth point cloud estimation and the preset constraint condition, wherein the preset constraint condition comprises one or more of epipolar geometric constraint, brightness constraint, gradient constraint and smoothness constraint. In a preferred embodiment of the present invention, the above four constraints can be included simultaneously, and the accurate depth point cloud is obtained by the following formula:
wherein, x: defining a pixel position (x, y) in the reference view angle c image, and its brightness is defined as i (x); x is the number ofb：＝(xb，ybC) is the epipolar point at view c +1, w is the offset of the corresponding point of x at view c + 1;
is a spatial gradient operator; β (x) is the occlusion map, 1 for pixels of the non-occluded area, and 0 otherwise. Considering the influence of the outliers in the model hypothesis, we use a robust penalty function
To produce a total variation regularization where ε is a small value (set to 0.001 in the experiment) the formula includes four constraints: epipolar geometric constraint (x)b+ d ═ x + w), luminance constraint (I (x)b+ d) ═ i (x)), gradient constraints
And smoothness constraints
And step S104, fusing the obtained depth point clouds of all the visual angles to obtain a static three-dimensional model. The method specifically comprises the following steps:
step S301, the depth point clouds of all the visual angles are fused, and some field values are removed through contour constraint.
And S302, reconstructing a complete surface model by a moving cube method to obtain a static three-dimensional model.
The method can ensure the accuracy and the integrity of the reconstructed shape of the static three-dimensional model, and the accuracy and the integrity of the static three-dimensional model are the basis of the reconstruction of the dynamic three-dimensional model.
As shown in fig. 3, a flowchart of a method for capturing and reconstructing a dynamic three-dimensional model according to an embodiment of the present invention includes the following steps:
step S401, convert the surface model of the static three-dimensional model into a body model, and represent it as a default scene for motion tracking.
Step S402, obtaining the initial three-dimensional movement of the model vertex at the next moment. Specifically, the following steps may be included:
in step S501, the optical flow of each view angle image at the next time is calculated.
In step S502, a scene stream of visible points is obtained from each perspective optical flow and the adjacent perspective optical flows, and a relatively large value, for example 10000, is assigned to a scene stream of invisible points.
Step S503, using the obtained scene flow of each view angle as a column, constructing a matrix M e im×nWherein m is the number of surface vertices.
Step S504, based on the sparse representation theory, a new matrix X is obtained by solving the following low-rank matrix recovery problem.
minimize ||X||*
Wherein X is an unknown variable and Ω is [ m ]]×[n]A subset of the complete set of elements ([ n ]]Defined as the sequence of numbers 1, K, n),
for the sampling operator, define as
Step S505, the average value of each row in the matrix X is taken as the movement of the vertex corresponding to the row
Thereby obtaining the vertex position of the next moment
And S403, selecting an accurate vertex from the acquired vertices as a position constraint of the body deformation according to a preset space-time constraint condition. In this embodiment of the present invention, the predetermined space-time constraint condition includes:
wherein, Psil n(v′i) Is the profile error of the estimated value, if v'iIf the pixel point projected to the n image of the camera at the next moment is in the outline, the function value is 1, otherwise, the function value is 0; v (i) is viA set of visible cameras; n is a radical ofvIs the number of visible cameras; pz n(p(vi)，p(v′i) Calculating viAnd v'iZNCC correlation between projected locations on camera n images; n is a radical ofsIs a vertex viThe number of direct neighbors.
And S404, driving the Laplace body deformation frame to update the dynamic three-dimensional model according to the position constraint. Specifically, the method comprises the following steps:
step S601, establishing a Laplace body deformation linear system for each v'iIs provided with
Wherein R isiAnd RjIs a rotation matrix and is initialized to a unit matrix.
Step S602, defining a covariance matrix
To CiSingular value decomposition is carried out by
Then
If det (R)i) If the value is less than or equal to 0, changing UiThe symbol of the column corresponding to the smallest singular value;
step S603, if the contour error is smaller than a given threshold, the model is updated, otherwise, the step S601 is returned to.
As a preferred embodiment of the present invention, the above-mentioned capturing and reconstructing method for a static three-dimensional model and a dynamic three-dimensional model of the present invention can be used simultaneously, as shown in fig. 4, which is a schematic block diagram of the entire dynamic three-dimensional reconstructing method according to the embodiment of the present invention.
Fig. 5 shows the results of a dynamic three-dimensional model obtained by applying the proposed inventive method to two long-term sequences. The first graph of each sequence result is a general graph formed by putting models at various moments together, and the subsequent graphs are the modeling results at various moments respectively.
The embodiment of the invention also provides a system for capturing and reconstructing the static three-dimensional model, which comprises: a plurality of cameras surrounding the annular field and a static three-dimensional model reconstruction device. The plurality of cameras surrounding the annular field are used for acquiring images of a moving object in the annular field; the static three-dimensional model reconstruction device is used for acquiring a visual shell model, acquiring depth point clouds of all visual angles according to images of all visual angles, the visual shell model and preset constraint conditions, and fusing the acquired depth point clouds of all visual angles to obtain a static three-dimensional model. The specific working process of the static three-dimensional model reconstruction device may refer to the above embodiments of the method for capturing and reconstructing a static three-dimensional model, and will not be described herein again.
In addition, the embodiment of the invention also provides a system for capturing and reconstructing the dynamic three-dimensional model,
the method comprises the following steps: a plurality of cameras surrounding the annular field, a static three-dimensional model acquisition device and a dynamic three-dimensional model reconstruction device. The plurality of cameras surrounding the annular field are used for acquiring images of a moving object in the annular field; the static three-dimensional model obtaining device is used for obtaining a static three-dimensional model; and the dynamic three-dimensional model reconstruction device is used for converting the surface model of the static three-dimensional model into a body model, representing the body model as a default scene of motion tracking, acquiring initial three-dimensional motion of model vertexes at the next moment, selecting an accurate vertex from the acquired vertexes as position constraint of body deformation according to a preset space-time constraint condition, and driving the Laplace body deformation framework to update the dynamic three-dimensional model according to the position constraint. The specific working process of the static and dynamic three-dimensional model reconstruction device may refer to the above embodiments of the method for capturing and reconstructing the static and dynamic three-dimensional models, and will not be described herein again.
The invention can ensure the accuracy and the integrity of the reconstructed shape of the static three-dimensional model, and in addition, the invention designs a new three-dimensional motion estimation method based on a sparse representation theory and a deformation optimization frame based on a body model, thereby obtaining a high-quality dynamic reconstruction result. In addition, the invention can be independent of three-dimensional scanners and optical markers, thus having low cost and being capable of tracking the movement of people wearing any dress.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (6)
1. A method for capturing and reconstructing a static three-dimensional model is characterized by comprising the following steps:
carrying out image acquisition on a moving object in the annular field;
acquiring a visual shell model;
obtaining depth point clouds of all the visual angles according to the images of all the visual angles, the visual shell model and preset constraint conditions;
fusing the obtained depth point clouds of all the visual angles to obtain a static three-dimensional model,
wherein,
the obtaining of the depth point cloud of each view angle according to each view angle image, the visual shell model and the preset constraint condition comprises:
intersecting the images of all the visual angles with the obtained visual shell model to obtain visible point clouds of all the visual angles;
projecting the visible point cloud of each visual angle to the visual angle image to obtain initial depth point cloud estimation;
obtaining an accurate depth point cloud according to the initial depth point cloud estimation and the preset constraint condition;
the preset constraint conditions comprise one or more of epipolar geometric constraint, brightness constraint, gradient constraint and smoothness constraint;
obtaining an accurate depth point cloud by the following formula:
wherein,d(a, b, 1) is an initial depth point cloud estimate,x: where (x, y, c) is a pixel position (x, y) in the reference view angle c image, and i (x) is the luminance of the pixel position;x b ：＝(xb，yband c) is an epipolar point at the viewing angle c + 1;
is a spatial gradient operator; beta (x) is an occlusion map,
is a robust penalty function.
2. The method for capturing and reconstructing a static three-dimensional model according to claim 1, wherein the step of fusing the obtained depth point clouds of the respective view angles to obtain the static three-dimensional model comprises:
fusing the depth point clouds of all the visual angles and removing field values through contour constraint; and
and reconstructing a complete surface model by a moving cube method to obtain a static three-dimensional model.
3. The method for capturing and reconstructing a static three-dimensional model according to claim 1 or 2, further comprising:
and constructing a dynamic three-dimensional model according to the static three-dimensional model.
4. The method for capturing and reconstructing a static three-dimensional model as claimed in claim 3, wherein said constructing a dynamic three-dimensional model from said static three-dimensional model comprises:
converting a surface model of the static three-dimensional model into a body model, and using the body model as a default scene representation of motion tracking;
acquiring initial three-dimensional motion of a model vertex at the next moment;
selecting an accurate vertex from the obtained vertexes as position constraint of body deformation according to a preset space-time constraint condition;
driving the Laplace body deformation frame to update a dynamic three-dimensional model according to the position constraint,
wherein,
the obtaining of the initial three-dimensional motion of the model vertex at the next moment comprises:
calculating the optical flow of each visual angle image at the next moment;
obtaining a scene stream of visible points from each visual angle optical stream and adjacent visual angle optical streams, and assigning a relatively large value to the scene stream of invisible points;
constructing a matrix M e i by taking the obtained scene flow of each visual angle as a columnm×nWherein m is the number of surface peaks;
obtaining a matrix X based on a sparse representation theory;
taking the average value of each row in the matrix X as the movement of the vertex corresponding to the row
Thereby obtaining the vertex position of the next moment
Wherein,
the obtaining of the matrix X based on the sparse representation theory comprises:
a new matrix X is obtained by solving the following low rank matrix recovery problem,
minimize ||X||*
wherein X is an unknown variable and Ω is [ m ]]×[n]A subset of the complete set of elements, [ n ]]Defined as a sequence of numbers 1, n,
for the sampling operator, define as
The predetermined space-time constraints include:
wherein,
is the profile error of the estimated value, if v'iIf the pixel point projected to the n image of the camera at the next moment is in the outline, the function value is 1, otherwise, the function value is 0;
is viA set of visible cameras; n is a radical ofvIs the number of visible cameras;calculating viAnd v'iZNCC correlation between projected locations on camera n images; n is a radical ofsIs a vertex viThe number of direct neighbors.
5. The method for capturing and reconstructing a static three-dimensional model according to claim 4, wherein said updating the dynamic three-dimensional model based on the position constraint-driven Laplace volume deformation framework comprises:
initializing a rotation matrix as an identity matrix, Ri＝Rj＝I；
Optimizing by using Laplace body deformation;
obtaining a new rotation matrix RiAnd Rj；
And judging whether the contour error is smaller than a preset value, if so, updating the dynamic three-dimensional model, and if not, continuing to perform optimization by using the Laplace body deformation.
6. A system for capturing and reconstructing a static three-dimensional model, comprising:
the cameras surround the annular field and are used for acquiring images of a moving object in the annular field; and
a static three-dimensional model reconstruction device, which is used for obtaining a visual shell model, obtaining depth point clouds of each visual angle according to each visual angle image, the visual shell model and a preset constraint condition, and fusing the obtained depth point clouds of each visual angle to obtain a static three-dimensional model,
wherein,
the static three-dimensional model reconstruction device intersects the images of all the visual angles with the obtained visible shell model to obtain visible point clouds of all the visual angles, projects the visible point clouds of all the visual angles to the images of the visual angles to obtain initial depth point cloud estimation, and obtains accurate depth point clouds according to the initial depth point cloud estimation and the preset constraint condition;
the preset constraint conditions comprise one or more of epipolar geometric constraint, brightness constraint, gradient constraint and smoothness constraint;
the static three-dimensional model reconstruction device obtains accurate depth point cloud through the following formula:
wherein,d(a, b, 1) is an initial depth point cloud estimate,x: where (x, y, c) is a pixel position (x, y) in the reference view angle c image, and i (x) is the luminance of the pixel position;x b ：＝(xb，yband c) is an epipolar point at the viewing angle c + 1;
is a spatial gradient operator; beta (x) is an occlusion map,
is a robust penalty function.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010101411826A CN101833786B (en) | 2010-04-06 | 2010-04-06 | Method and system for capturing and rebuilding three-dimensional model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010101411826A CN101833786B (en) | 2010-04-06 | 2010-04-06 | Method and system for capturing and rebuilding three-dimensional model |
Related Child Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 201110167593 Division CN102222361A (en) | 2010-04-06 | 2010-04-06 | Method and system for capturing and reconstructing 3D model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101833786A CN101833786A (en) | 2010-09-15 |
| CN101833786B true CN101833786B (en) | 2011-12-28 |
Family
ID=42717847
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2010101411826A Active CN101833786B (en) | 2010-04-06 | 2010-04-06 | Method and system for capturing and rebuilding three-dimensional model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101833786B (en) |
Families Citing this family (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102306390B (en) * | 2011-05-18 | 2013-11-06 | 清华大学 | Method and device for capturing movement based on framework and partial interpolation |
| AU2011203028B1 (en) * | 2011-06-22 | 2012-03-08 | Microsoft Technology Licensing, Llc | Fully automatic dynamic articulated model calibration |
| CN102446366B (en) * | 2011-09-14 | 2013-06-19 | 天津大学 | Time-space jointed multi-view video interpolation and three-dimensional modeling method |
| CN102722908B (en) * | 2012-05-25 | 2016-06-08 | 任伟峰 | Method for position and device are put in a kind of object space in three-dimension virtual reality scene |
| CN102800127B (en) * | 2012-07-18 | 2014-11-26 | 清华大学 | Light stream optimization based three-dimensional reconstruction method and device |
| CN103903300A (en) * | 2012-12-31 | 2014-07-02 | 博世汽车部件(苏州)有限公司 | Object surface height reconstructing method, object surface height reconstructing system, optical character extracting method and optical character extracting system |
| CN103927787A (en) * | 2014-04-30 | 2014-07-16 | 南京大学 | Method and device for improving three-dimensional reconstruction precision based on matrix recovery |
| CN104599314A (en) * | 2014-06-12 | 2015-05-06 | 深圳奥比中光科技有限公司 | Three-dimensional model reconstruction method and system |
| CN105488823B (en) * | 2014-09-16 | 2019-10-18 | 株式会社日立制作所 | CT image rebuilding method, CT equipment for reconstructing image and CT system |
| US20160140733A1 (en) * | 2014-11-13 | 2016-05-19 | Futurewei Technologies, Inc. | Method and systems for multi-view high-speed motion capture |
| US10127709B2 (en) * | 2014-11-28 | 2018-11-13 | Panasonic Intellectual Property Management Co., Ltd. | Modeling device, three-dimensional model generating device, modeling method, and program |
| CN107170037A (en) * | 2016-03-07 | 2017-09-15 | 深圳市鹰眼在线电子科技有限公司 | A kind of real-time three-dimensional point cloud method for reconstructing and system based on multiple-camera |
| CN108140252A (en) * | 2016-09-08 | 2018-06-08 | 深圳市大富网络技术有限公司 | A kind of generation method and relevant device of square animation |
| US10572720B2 (en) * | 2017-03-01 | 2020-02-25 | Sony Corporation | Virtual reality-based apparatus and method to generate a three dimensional (3D) human face model using image and depth data |
| CN107358645B (en) * | 2017-06-08 | 2020-08-11 | 上海交通大学 | Product three-dimensional model reconstruction method and system |
| TWI657407B (en) * | 2017-12-07 | 2019-04-21 | 財團法人資訊工業策進會 | Three-dimensional point cloud tracking apparatus and method by recurrent neural network |
| CN108769361B (en) * | 2018-04-03 | 2020-10-27 | 华为技术有限公司 | Control method of terminal wallpaper, terminal and computer-readable storage medium |
| CN109271893B (en) * | 2018-08-30 | 2021-01-01 | 百度在线网络技术（北京）有限公司 | Method, device, equipment and storage medium for generating simulation point cloud data |
| WO2021051220A1 (en) * | 2019-09-16 | 2021-03-25 | 深圳市大疆创新科技有限公司 | Point cloud fusion method, device, and system, and storage medium |
| CN112001958B (en) * | 2020-10-28 | 2021-02-02 | 浙江浙能技术研究院有限公司 | Virtual point cloud three-dimensional target detection method based on supervised monocular depth estimation |
| WO2022087932A1 (en) * | 2020-10-29 | 2022-05-05 | Huawei Technologies Co., Ltd. | Non-rigid 3d object modeling using scene flow estimation |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6791542B2 (en) * | 2002-06-17 | 2004-09-14 | Mitsubishi Electric Research Laboratories, Inc. | Modeling 3D objects with opacity hulls |
| CN100557640C (en) * | 2008-04-28 | 2009-11-04 | 清华大学 | A kind of interactive multi-vision point three-dimensional model reconstruction method |
| CN101650834A (en) * | 2009-07-16 | 2010-02-17 | 上海交通大学 | Three dimensional reconstruction method of human body surface under complex scene |
-
2010
- 2010-04-06 CN CN2010101411826A patent/CN101833786B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN101833786A (en) | 2010-09-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101833786B (en) | Method and system for capturing and rebuilding three-dimensional model | |
| CN102222361A (en) | Method and system for capturing and reconstructing 3D model | |
| CN107578436B (en) | Monocular image depth estimation method based on full convolution neural network FCN | |
| De Aguiar et al. | Marker-less deformable mesh tracking for human shape and motion capture | |
| CN108711185B (en) | Three-dimensional reconstruction method and device combining rigid motion and non-rigid deformation | |
| Dou et al. | Scanning and tracking dynamic objects with commodity depth cameras | |
| Hoppe et al. | Online Feedback for Structure-from-Motion Image Acquisition. | |
| EP2595116A1 (en) | Method for generating depth maps for converting moving 2d images to 3d | |
| CN104915978B (en) | Realistic animation generation method based on body-sensing camera Kinect | |
| CN109242873A (en) | A method of 360 degree of real-time three-dimensionals are carried out to object based on consumer level color depth camera and are rebuild | |
| Xu et al. | 3d virtual garment modeling from rgb images | |
| JP4761670B2 (en) | Moving stereo model generation apparatus and method | |
| Sizintsev et al. | Spatiotemporal stereo and scene flow via stequel matching | |
| CN115951784B (en) | Method for capturing and generating motion of wearing human body based on double nerve radiation fields | |
| Li et al. | Animated 3D human avatars from a single image with GAN-based texture inference | |
| CN106683181A (en) | Method for reconstructing three-dimensional human body dense surface motion field | |
| Li et al. | Three-dimensional motion estimation via matrix completion | |
| Remondino et al. | 3D reconstruction of human skeleton from single images or monocular video sequences | |
| CN113920270A (en) | Layout reconstruction method and system based on multi-view panorama | |
| CN110490973B (en) | Model-driven multi-view shoe model three-dimensional reconstruction method | |
| CN112132971A (en) | Three-dimensional human body modeling method, device, electronic equipment and storage medium | |
| De Aguiar et al. | Marker-less 3D feature tracking for mesh-based human motion capture | |
| Mahmoud et al. | Fast 3d structure from motion with missing points from registration of partial reconstructions | |
| Remondino et al. | Human motion reconstruction and animation from video sequences | |
| JP2009048305A (en) | Shape analysis program and shape analysis apparatus |
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 |