CN105913493B - A kind of interaction attributes based on single image and threedimensional model joint mapping method - Google Patents
A kind of interaction attributes based on single image and threedimensional model joint mapping method Download PDFInfo
- Publication number
- CN105913493B CN105913493B CN201610261400.7A CN201610261400A CN105913493B CN 105913493 B CN105913493 B CN 105913493B CN 201610261400 A CN201610261400 A CN 201610261400A CN 105913493 B CN105913493 B CN 105913493B
- Authority
- CN
- China
- Prior art keywords
- component
- image
- interaction attributes
- threedimensional model
- model
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
- G06T17/10—Constructive solid geometry [CSG] using solid primitives, e.g. cylinders, cubes
Landscapes
- Physics & Mathematics (AREA)
- Geometry (AREA)
- Engineering & Computer Science (AREA)
- Computer Graphics (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Processing Or Creating Images (AREA)
Abstract
The present invention provides a kind of interaction attributes based on single image and threedimensional model joint mapping method, includes the following steps: to input individual image with interaction attributes object and complete the label of object bounding box and component；It establishes the structural model of the object and completes part dimension, layout optimization；Active force label is done to all parts on structural model, constructs component stress graph structure；It determines each unit type, generates the threedimensional model with interaction attributes；Iteration above step completes image object internal structure design.The present invention overcomes the limitation that object internal structural information can not be learnt from image, with allowing user individual designed image object internal structure, to obtain a complete threedimensional model with interaction attributes.This method is not only suitable for the professional users such as designer, is also applied for common amateur user, has broad application prospects in fields such as threedimensional model Rapid Prototype Design, interior decoration design and e-commerce.
Description
Technical field
The present invention relates to field of Computer Graphics, specifically a kind of interaction attributes based on single image and three-dimensional
Model joint mapping method.
Background technique
Three-dimensional modeling always is one of main direction of studying of field of Computer Graphics.Interaction attributes and threedimensional model
Joint mapping can quickly generate the threedimensional model that user can directly interact, therefore handed over for common having in daily life
The object of mutual attribute, it is a very interesting and significant job that building, which generates the threedimensional model with interaction attributes, three
The fields such as dimension module Rapid Prototype Design, interior decoration design and e-commerce have broad application prospects.
In field of Computer Graphics, the three-dimensional modeling based on image has had long-term development and research, and existing
It is being still the hot spot of research, many researchers are dedicated to the research of the three-dimensional modeling based on image.Byong Mok Oh etc.
People proposed a kind of modeling based on image and editing system in 2001.By being layered to input picture, reconstruct three-dimensional scenic,
Allow to direct editing input picture.Nianjuan Jiang et al. proposed a kind of having based on single image in 2009
The three-dimensional modeling method of texture needs user's interaction guidance in modeling process.Manfred Lau et al. proposed one in 2010
A frame allows user using image as a rough guidance, designs threedimensional model personalizedly.However, being defeated with image
Entering to construct threedimensional model is still a challenging task.Since different types of model has different structure and spy
Sign, so being hardly formed a kind of unified modeling method is suitable for different types of model construction.Therefore, it is based on image at this stage
Three-dimensional modeling research all only consider a certain certain types of object, such as the Plants modeling based on image, based on image
City Modeling etc..
In terms of the object three-dimensional modeling with interaction attributes, also there are some researchs and progress.Wanbin Pan
Et al. a kind of furniture modeling method based on sample was proposed in 2011, by reusing existing threedimensional model, and in model
Upper addition parameter limitation generates new three-dimensional furniture model to construct.Youyi Zheng et al. proposed one kind in 2013
The reasonable model of method of geometry constructing function based on symmetry substructure matching, replacement.Manfred Lau et al. is 2011
It proposes a kind of method year and structural analysis is carried out to the three-dimensional furniture model of input, isolate all parts, and generate each portion
Connection structure between part, so that the model generated has interaction attributes.Yuliy Schwartzburg et al. was mentioned in 2013
A kind of calculating design method is gone out, can have been assembled by " chain flat components " and generate threedimensional model.This modeling method can expand
Open up the assembling building of mock-up funiture.Yuliang Rong et al. proposed a kind of from individual RGBD image recovery function in 2016
The interactive approach of energy property prototype, this method has had to believable functional parts, but there is no modeling generation is actual
Threedimensional model.
The above present Research shows currently based on the three-dimensional modeling method of image mostly just for certain types of right
As needing to be goed deep into for the modeling of the object with interaction attributes, and existing modeling method seldom relates to object interaction
Attribute and threedimensional model joint mapping problem.
Summary of the invention
The technology of the present invention solves the problems, such as: overcoming the shortcomings of existing methods, the present invention provides a kind of based on single image
Interaction attributes and threedimensional model joint mapping method, this method have the image of interaction attributes object as input, lead to using single width
The simple marking of user is crossed, rapid build generates image object 3 d structure model.Then, it is used on image object structural model
Family carries out the label of component stress condition again, constructs the stress graph structure of all parts, so that it is determined that the type of all parts, friendship
Mutual mode and connection structure.Finally, generating the threedimensional model with interaction attributes using preset basic element of character model.
To complete goal of the invention, the technical solution adopted by the present invention is that: a kind of interaction attributes and three based on single image
Dimension module joint mapping method, method includes the following steps:
Step 1: inputting individual image with interaction attributes object and completes the label of object bounding box and component；
Step 2: establishing the structural model of image object and completes part dimension, layout optimization；
Step 3: doing active force label to all parts on structural model, constructs component stress graph structure；
Step 4: determining each unit type, generates the threedimensional model with interaction attributes；
Step 5: iteration above step completes image object internal structure design.
Wherein, at " interaction attributes " described in step 1, refer to the component interactive mode based on translating, rotate, such as
The pull of drawer, the rotation etc. of door and window." structural model of image object " described in step 2, refers to image object
Main body frame is abstracted as a cube, and the component of object is abstracted as to some cubes for being attached to main body frame solid front faces
Thin piece." part dimension, layout optimization " described in step 2 refers to since user hand is used force and slightly marked in step 1
Component may in size, layout misalignment, so need to consider the size and location relationship of all parts component adjacent thereto,
Each component is aligned.Each component side length needs to meet:
Wherein, SiIndicate the length on i-th side after optimizing, IiIndicate the length on i-th side of user's mark, SijIt indicates
With i-th set while adjacent all, n indicates the quantity on side to be optimized, and λ is the parameter of balance front and back two.
Wherein, " the component stress graph structure " described in step 3 refers to the stress surface of a component is (i.e. one long
Square face) each side be abstracted as a node, i.e., there are four side node, they are sequentially connected, and furthermore add on the stress surface
Add a face node, on each node can one active force of label, the node that active force is marked is stress node, thus
The structure of composition is " component stress graph structure ".By the graph structure, the concrete type of the component, different type can be determined
Component have different motion mode and connection structure, thus can obtain the three-dimensional mould with interaction attributes in step 4
Type.
Wherein, " internal structure design " described in step 5 is because the object on input picture is typically all to close
Conjunction state, it is any structure on earth inside object that user can not learn from image.Therefore, the present invention allows user above-mentioned
Its internal structure is further designed on the basis of model.
Compared with prior art, beneficial feature is the present invention:
1, the present invention uses individual image with interaction attributes object as input, can be achieved under user's interaction guidance
The interaction attributes and threedimensional model joint mapping of image object.
2, the invention proposes a kind of interactive boot can interactive component modeling method so that generate threedimensional model portion
Part has interaction attributes.
3, the present invention allows user to do personalized internal structure design to image object model.
Detailed description of the invention
Fig. 1 is the method for the present invention implementation flow chart；
Fig. 2 is the method for the invention flow diagram；
Fig. 3 is component stress figure structure schematic representation of the invention；
Fig. 4 is four kinds of standarized component models that the present invention defines；
Fig. 5 is internal structure design flow diagram of the invention.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, lower mask body is so that furniture models as an example,
In conjunction with attached drawing, explanation is explained in detail to method of the invention.
The invention proposes interaction attributes and threedimensional model joint mapping method based on single image.It is modeled as with furniture
Example, the method as inputting, go out the bounding box of furniture and each by the rough hand labeled of user using individual furniture image
Component, can rapid build generate the structural model of furniture, which is abstracted as one for the component of furniture and is attached to encirclement
Thin piece of the cube of box solid front faces, while user's mark component, this method intercepted the screenshot of all parts so as to
The mock-up funiture with texture mapping can be generated.Further it is proposed that a kind of interactive boot can interactive component build
Mould method.User marks the stress condition of all parts on above structure model, and this method is by establishing component force diagram knot
Structure determines the concrete type of furniture parts, generate have can interactive component mock-up funiture.Due to that can not learn house from image
The internal structure of tool, the present invention design furniture internal structure with also allowing user individual.The present invention is in a small amount of user interaction
Under guidance, can mock-up funiture that quickly and conveniently generating unit can interact, be suitable for professional furniture design teacher, be also applied for
Common amateur user.
The present invention is based on the interaction attributes of single image and threedimensional model joint mapping method, operating processes such as figure, 2 institutes
Show, specific implementation step is as follows:
Step 1: inputting individual furniture image and completes the label of furniture bounding box and all parts.
The present invention carries out furniture modeling using individual furniture image as input.Input picture is carried out firstly the need of user
Label, label are divided into two kinds: to the label of furniture bounding box, and to the label of furniture all parts.Such as (a) institute in Fig. 2
Show, the cube that dotted line is constituted is the bounding box of furniture, and solid line quadrangle is all parts of furniture.To furniture bounding box
The label present invention use existing method, user marks six vertex of the furniture outermost on the image, can from
The dynamic coordinate that other two vertex is calculated, to identify its corresponding cube.Then, in above-mentioned bounding box cube
Front, user completes the label to furniture parts by marking the quadrangle borderline regions of all parts.
Step 2: establishing furnature construction model and completes part dimension, layout optimization.
By the label of step 1, the present invention can quickly generate the structural model of furniture.So-called furnature construction model is such as schemed
Shown in (b) in 2, refers to and the main body frame of furniture is abstracted as a cube, the component of furniture is abstracted as and is attached to main body
Thin piece of some cubes of frame solid front faces.Further, since in step 1 user hand use force the component that slightly marks may be
Misalignment in size, layout, so the invention proposes a kind of component automatic aligning algorithms to be aligned adjacent component.
The algorithm needs to consider that the neighbouring relations of component and the side length of all parts, each side length need to meet:
Wherein, SiIndicate the length on i-th side after optimizing, IiIndicate the length on i-th side of user's mark, SijIt indicates
With i-th set while adjacent all, n indicates the quantity on side to be optimized, and λ is the parameter of balance front and back two.Above-mentioned public affairs
Formula is a quadratic function, can quickly acquire result by conjugate gradient algorithms.
Step 3: doing active force label to all parts on structural model, constructs component stress graph structure.
On the basis of above-mentioned furnature construction model, the invention proposes a kind of interactive boot can interactive component modeling
Method.Stress condition is marked to each component in structural model by user, component stress graph structure is constructed, determines each component
Type.In order to facilitate user's mark, the present invention introduces five " control points " to each furniture parts, in (b) in Fig. 2
Five black small cubes, four dominating pair of vertices of surrounding answer the four edges of furniture parts, record the stress condition on each side, center control
The face of the corresponding furniture parts of system point, records the stress condition in the face.Each control point can mark the active force in certain direction, such as Fig. 2
In (b) shown in arrow.After the label for completing component stress condition, the present invention corresponds to each component construction force diagram knot
Structure, as shown in Figure 3.Including the node of three types in component stress graph structure: side node, face node and stress node.
While node have two attribute: at a distance from stress node, this while node corresponding side and Impact direction angle.Face node is only
It can be needed individually to handle by the power in the vertical face.As shown in figure 3, node a is then a stress node in (b) in Fig. 3,
Its Impact direction is that horizontally to the right, node a is also a stress node in (c) in Fig. 3, and Impact direction is vertical outside.
Step 4: determining each unit type, generate have can interactive component mock-up funiture.
On the basis of above-mentioned component stress graph structure, in order to determine the concrete type and motion mode of all parts,
The present invention is for each node V in component stress graph structureiParameter (d is seti,θi), wherein diIt is the node apart from stress knot
Point distance,θiThe angle for referring to the active force of the corresponding side of the node and user's mark, since face node can only be by perpendicular to this
The active force in face, so needing individually processing.Type and the position of member connection structure are determined eventually by following formula.
Wherein V ' is the set for needing to be arranged the node of connection structure.If having θ for all ii≠ 0, then the company
Binding structure is hinge types, and position is at the node farthest apart from stress node, and otherwise, connection structure is then sliding type, even
Binding structure, that is, sliding slot position is at the node parallel with user's mark force direction.In addition, if the effect of user's mark
Power is located on a face node, then the member connection structure is the sliding type along force direction.There are also a kind of special feelings
Condition, i.e., the component stress graph structure of user's mark useless, such case present invention then default component not movable herein, i.e.,
For space.
Invention defines four kinds of most common furniture parts types: revolving door, sliding door, drawer and hollow grid,
And the master pattern of these four components is given, as shown in Figure 4.
On the basis of the above, the present invention cuts Furniture Body cube using the method for constructive solid geometry, right
Above-mentioned standard component is deformed, is translated, and assembles them into a complete mock-up funiture, and according to the concrete type of component and
Motion mode adds kinematic constraint to the component in mock-up funiture, the mock-up funiture that component can interact is finally obtained, in Fig. 2
(c) shown in.
Step 5: iteration above step completes the design of inside of furniture structure.
Due to that can not obtain inside of furniture structure from furniture image, the present invention designs furniture with allowing user individual
Internal structure, operating process are as shown in Figure 5.(a) in Fig. 2, which is used, to be included the following steps: to the design of inside of furniture structure
On the basis of the label of family, by user's mark subassembly, structural model is established, by marking stress condition on structural model,
Subassembly type is determined, to generate final internal structure model.Internal structure design is completed, final family can be obtained
Has model, as shown in (d) in Fig. 2.The component textures intercepted in step 1 are attached on mock-up funiture, band has just been obtained
The mock-up funiture of textured textures, as shown in (e) in Fig. 2.So far to get to one with textures, have can interact portion
Part, the complete furniture threedimensional model with internal structure.
The foregoing is merely some basic explanations of the invention, any equivalent change that technical solution according to the present invention is done
It changes, is within the scope of protection of the invention.
Claims (5)
1. a kind of interaction attributes based on single image and threedimensional model joint mapping method, it is characterised in that including following step
It is rapid:
Step (1) inputs individual with interaction attributes object images and completes the label of object images bounding box and component；
Step (2) establishes the structural model of image object and completes part dimension, layout optimization；
Step (3) does active force label to all parts on structural model, constructs component stress graph structure；
Step (4) determines each unit type according to component stress graph structure, generates the threedimensional model with interaction attributes；
Step (5) iteration above step, completes the internal structure design of image object；
Part dimension described in the step (2), layout optimization are: in step (1) user hand use force the component slightly marked can
Can in size, layout misalignment, need to consider the size and location relationship of all parts component adjacent thereto, so that each component
It can be aligned, each component side length needs to meet:
Wherein, SiIndicate the length on i-th side after optimizing, IiIndicate the length on i-th side of user's mark, ∑ SijIndicate with
I-th set while adjacent all, n indicate the quantity on side to be optimized, and λ is the parameter of balance front and back two；
Each unit type of determination described in the step (4), generate have interaction attributes threedimensional model, refer to according to component by
The information for trying hard to the position of stress node in structure, suffered active force, determine the concrete types of all parts, motion mode, with
And the position of connection structure, the component in threedimensional model thus generated are just provided with the attribute that can be interacted.
2. the interaction attributes according to claim 1 based on single image and threedimensional model joint mapping method, feature
Be: interaction attributes described in step (1) refer to the component interactive mode based on translating, rotate, the pull including drawer, door
The rotation of window.
3. the interaction attributes according to claim 1 based on single image and threedimensional model joint mapping method, feature
Be: the configuration model of step (2) described image object is: the main body frame of image object is abstracted as one cube
The all parts of image object are abstracted as thin piece of some cubes for being attached to main body frame solid front faces, are stood by these by body
The model that cube is constituted is the structural model of image object.
4. the interaction attributes according to claim 1 based on single image and threedimensional model joint mapping method, feature
Be: the building of component stress graph structure described in step (3) be by the stress surface of a component, i.e. rectangle face it is each
A side is abstracted as a node, and there are four side nodes altogether, they are sequentially connected, and a face knot is furthermore added on the stress surface
Point, one active force of label on each node, the node that active force is marked is stress node, and the structure thus constituted is
Component stress graph structure.
5. the interaction attributes according to claim 1 based on single image and threedimensional model joint mapping method, feature
Be: image object internal structure design described in step (5) is: can not obtain object internal structural information from image, permit
Perhaps user individual designed image object internal structure, user marks on the profile cube in object interior design region
Subassembly establishes structural model, by marking stress condition on structural model, determines concrete type, the movement side of subassembly
The position of formula and connection structure, to generate final internal structure model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610261400.7A CN105913493B (en) | 2016-04-25 | 2016-04-25 | A kind of interaction attributes based on single image and threedimensional model joint mapping method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610261400.7A CN105913493B (en) | 2016-04-25 | 2016-04-25 | A kind of interaction attributes based on single image and threedimensional model joint mapping method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105913493A CN105913493A (en) | 2016-08-31 |
| CN105913493B true CN105913493B (en) | 2019-03-08 |
Family
ID=56752793
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610261400.7A Active CN105913493B (en) | 2016-04-25 | 2016-04-25 | A kind of interaction attributes based on single image and threedimensional model joint mapping method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105913493B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102508991A (en) * | 2011-09-30 | 2012-06-20 | 北京航空航天大学 | Method of constructing virtual experiment teaching scene based on image material |
| CN102508994A (en) * | 2011-09-30 | 2012-06-20 | 北京航空航天大学 | Virtual experiment teaching oriented dynamic component visualization method |
| CN103279599A (en) * | 2013-05-15 | 2013-09-04 | 西北工业大学 | Product modeling method based on interactive feature pair |
-
2016
- 2016-04-25 CN CN201610261400.7A patent/CN105913493B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102508991A (en) * | 2011-09-30 | 2012-06-20 | 北京航空航天大学 | Method of constructing virtual experiment teaching scene based on image material |
| CN102508994A (en) * | 2011-09-30 | 2012-06-20 | 北京航空航天大学 | Virtual experiment teaching oriented dynamic component visualization method |
| CN103279599A (en) * | 2013-05-15 | 2013-09-04 | 西北工业大学 | Product modeling method based on interactive feature pair |
Non-Patent Citations (2)
| Title |
|---|
| Meta-representation of shape families;Noa Fish et al.;《ACM Transactions on Graphics》;20140731;第33卷(第4期);第1-11页 |
| Style-content separation by anisotropic part scales;Kai Xu et al.;《ACM Transactions on Graphics》;20101030;第29卷(第1期);第1-10页 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105913493A (en) | 2016-08-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Botsch et al. | Adaptive space deformations based on rigid cells | |
| Khalili | An XML-based approach for geo-semantic data exchange from BIM to VR applications | |
| Zhang et al. | Interactive tensor field design and visualization on surfaces | |
| Gao et al. | An approach to solid modeling in a semi-immersive virtual environment | |
| Jordao et al. | Crowd sculpting: A space‐time sculpting method for populating virtual environments | |
| CN109974703A (en) | A kind of construction method, system and the device of indoor navigation network | |
| Yun et al. | Using VR technology combined with CAD software development to optimize packaging design | |
| Museth et al. | Algorithms for interactive editing of level set models | |
| Wang et al. | Video-based fluid reconstruction and its coupling with SPH simulation | |
| Sun et al. | Sequential fusion of multi-view video frames for 3D scene generation | |
| CN105913493B (en) | A kind of interaction attributes based on single image and threedimensional model joint mapping method | |
| CN116229788A (en) | Design method, system and application method of virtual server assembly simulation training system | |
| Tan | Animation Image Art Design Mode Using 3D Modeling Technology | |
| CN107516341A (en) | Architectural wind environment method for visualizing based on AR | |
| Wu et al. | Freescup: A novel platform for assisting sculpture pose design | |
| Zhang et al. | Immersive product configurator for conceptual design | |
| Tang et al. | Performance-Based Generative Design, An investigation of the Parametric Nature of Architecture | |
| Zhang | Development of 3D Modeling Parameter Design in Furniture Design | |
| Chen et al. | View-aware image object compositing and synthesis from multiple sources | |
| Sookkaew et al. | Combining the Characteristics of the Buddha Statue from Photogrammetry and 3D Creation to Simulate the Combination of the Art of Creating Buddha Statue | |
| Kim et al. | OptCtrlPoints: Finding the Optimal Control Points for Biharmonic 3D Shape Deformation | |
| Chen | Research on mobile terminal sketch 3D modeling technology based on interactive design | |
| CN103729874B (en) | Flower modeling device and flower modeling method | |
| Chen et al. | Individualized Interior Design Combining CAD and Data Mining Technologies | |
| Xu et al. | Analysis of computer aided architecture and interior design based on building information modeling |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |