CN109304866A - The integrated equipment and method of 3D portrait are printed using self-service take pictures of 3D camera - Google Patents
The integrated equipment and method of 3D portrait are printed using self-service take pictures of 3D camera Download PDFInfo
- Publication number
- CN109304866A CN109304866A CN201811058666.7A CN201811058666A CN109304866A CN 109304866 A CN109304866 A CN 109304866A CN 201811058666 A CN201811058666 A CN 201811058666A CN 109304866 A CN109304866 A CN 109304866A
- Authority
- CN
- China
- Prior art keywords
- camera
- image
- coordinate
- depth
- printing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B29—WORKING OF PLASTICS; WORKING OF SUBSTANCES IN A PLASTIC STATE IN GENERAL
- B29C—SHAPING OR JOINING OF PLASTICS; SHAPING OF MATERIAL IN A PLASTIC STATE, NOT OTHERWISE PROVIDED FOR; AFTER-TREATMENT OF THE SHAPED PRODUCTS, e.g. REPAIRING
- B29C64/00—Additive manufacturing, i.e. manufacturing of three-dimensional [3D] objects by additive deposition, additive agglomeration or additive layering, e.g. by 3D printing, stereolithography or selective laser sintering
- B29C64/20—Apparatus for additive manufacturing; Details thereof or accessories therefor
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B29—WORKING OF PLASTICS; WORKING OF SUBSTANCES IN A PLASTIC STATE IN GENERAL
- B29C—SHAPING OR JOINING OF PLASTICS; SHAPING OF MATERIAL IN A PLASTIC STATE, NOT OTHERWISE PROVIDED FOR; AFTER-TREATMENT OF THE SHAPED PRODUCTS, e.g. REPAIRING
- B29C64/00—Additive manufacturing, i.e. manufacturing of three-dimensional [3D] objects by additive deposition, additive agglomeration or additive layering, e.g. by 3D printing, stereolithography or selective laser sintering
- B29C64/30—Auxiliary operations or equipment
- B29C64/386—Data acquisition or data processing for additive manufacturing
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B33—ADDITIVE MANUFACTURING TECHNOLOGY
- B33Y—ADDITIVE MANUFACTURING, i.e. MANUFACTURING OF THREE-DIMENSIONAL [3-D] OBJECTS BY ADDITIVE DEPOSITION, ADDITIVE AGGLOMERATION OR ADDITIVE LAYERING, e.g. BY 3-D PRINTING, STEREOLITHOGRAPHY OR SELECTIVE LASER SINTERING
- B33Y30/00—Apparatus for additive manufacturing; Details thereof or accessories therefor
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B33—ADDITIVE MANUFACTURING TECHNOLOGY
- B33Y—ADDITIVE MANUFACTURING, i.e. MANUFACTURING OF THREE-DIMENSIONAL [3-D] OBJECTS BY ADDITIVE DEPOSITION, ADDITIVE AGGLOMERATION OR ADDITIVE LAYERING, e.g. BY 3-D PRINTING, STEREOLITHOGRAPHY OR SELECTIVE LASER SINTERING
- B33Y40/00—Auxiliary operations or equipment, e.g. for material handling
Landscapes
- Chemical & Material Sciences (AREA)
- Engineering & Computer Science (AREA)
- Materials Engineering (AREA)
- Manufacturing & Machinery (AREA)
- Physics & Mathematics (AREA)
- Mechanical Engineering (AREA)
- Optics & Photonics (AREA)
- Image Processing (AREA)
Abstract
The invention belongs to 3D printing technique fields, the integrated equipment and method of 3D portrait are printed using self-service take pictures of 3D camera more particularly to a kind of, the equipment uses more 3D cameras, it is accurate to obtain human body color image and depth information data, by by the fusion of the image characteristic point of different perspectives, registration and splicing, quickly creating human 3d model, user can unrestricted choice print area, and pass through mobile payment, self-service completion 3D portrait printing.Integration degree is high for complete equipment, easy to operate, and visualized operation, measurement accuracy is higher, can the self-service whole process service for completing shooting, modeling, payment, printing in the case where unattended.
Description
Technical field
The invention belongs to 3D printing technique fields, print 3D portrait using self-service take pictures of 3D camera more particularly to a kind of
Integrated equipment and method.
Background technique
With universal, the mechanism that existing profession offer 3D portrait prints in the market of 3D printing technique.3D portrait prints most
It is important that the acquisition of human body three-dimensional data and human geometry's modeling, at present 3D portrait printing generally carry out people using 3D scanner
The capture and processing of volume data, process include scanning, repair figure, modeling, printing, and entire work flow needs to set using different types of
Standby instrument, not only data processing takes a long time, and user experience is bad, and it is high to also result in 3D portrait printing overall price.
Traditional 3D anthropometric scanning mode needs subscriber station before 3D body-scanner, keeps specific in certain time
Posture, scanner carry out processing using data of the 3 d modeling software to scanning after being scanned to human body and are converted to spatial point,
Virtual model is shown with point data.
It can also be used more set RGB cameras to carry out different angle to take pictures, then by 2D image identification method, from plane
In image, the geometrical model of 3 D human body is recovered by way of automatic or manual interaction.
Prior art disadvantage:
1,3D anthropometric scanning instrument majority is using the technologies such as structure light scan or laser scanning, but traditional 3 D human body
The personnel that scanning usually requires to receive professional training operate, and scanning speed is slower, and instrument cost is higher, is usually directed to
The specific posture of some of human body obtains human appearance data, therefore is difficult directly to meet people under different application environment
Demand.
2, since the 2D image of RGB camera shooting contains only the profile information of human body, the depth information without human body,
Depth information relies on pure software algorithm or special chip is calculated, therefore cannot obtain accurate 3D manikin.Meanwhile
Common RGB camera under dim environment and in the unconspicuous situation of feature and is not suitable for, and is illuminated by the light, object texture property shadow
It rings obviously, needs multi-angle of view data, it is bigger to the dependence of ancillary equipment.
Summary of the invention
Aiming at the problems existing in the prior art, the present invention, which provides, a kind of prints 3D portrait using self-service take pictures of 3D camera
Integrated equipment and method, integration degree is high for complete equipment, and visualized operation, easy to use, measurement accuracy is higher, can
Self-service completion shooting, models, payment, the whole process service of printing in the case where unattended.
To achieve the goals above, the following technical solution is employed by the present invention:
It is provided by the invention it is a kind of print the integrated equipment of 3D portrait using self-service take pictures of 3D camera, which includes:
More 3D cameras, for shooting human body color image and obtaining human depth's information；
Display unit is used for display operation order and image processing data；
Data processing unit, for the image information and depth information of the shooting of 3D camera described in calculation processing, and will not
With image characteristic point fusion, registration and the splicing at visual angle, human 3d model is quickly created；
3D printing unit, for printing the 3D portrait of desired target area；
Power supply unit, for providing electricity for more 3D cameras, display unit, data processing unit and 3D printing unit
Source.
The self-service method for printing 3D portrait of taking pictures of 3D camera is used the present invention also provides a kind of comprising the steps of:
3D camera is installed, more 3D cameras are both horizontally and vertically installed respectively, so that 3D camera is clapped
Take the photograph all spaces in range covering shooting area；
Print area is selected, before user takes pictures, first selects corresponding print area, different print areas on the display unit
The 3D camera of corresponding starting different number；
Selection payment, after selecting print area, the display unit of equipment generates payment interface, and user is swept by mobile terminal
Retouch the payment that payment two dimensional code completes printing expense；
Image Acquisition, 3D camera start work, carry out image data acquiring to human body with vertical direction from horizontal direction,
Image data includes the color image and depth information of human body；
Image registration and three-dimensional point cloud generate, for each of depth image pixel enclose corresponding RGB color from
And obtain the three-dimensional point cloud of body local；
Multi-view image splicing, since there is lap at every 3D camera visual angle, need to splice and be registrated under different perspectives
The body local point cloud that depth image represents, can be obtained complete 3 D human body point cloud data；
Human 3d model is constructed, Poisson resurfacing is carried out based on human body point cloud and obtains human 3d model；
3D portrait is printed, the 3D image data of region human body to be printed is transferred to 3D printing unit and beaten by data processing unit
Print 3D portrait.
Further, specific step is as follows for the installation 3D camera:
Due to 3D camera between object at a distance from it is closer, shooting effect is more accurate, to reach best shooting effect, takes pictures
Region is designed as circle, both horizontally and vertically installs more 3D cameras respectively；
If 3D camera horizontal direction field-of-view angle is α, 360 °/α platform 3D is horizontally mounted in the periphery of round shot region
Camera, 360 °/α platform 3D camera is to carry out 360 ° of horizontal direction to the human body being located at round shot region center every α
Image Acquisition；
In round shot region at interval of three 3D cameras of α right angle setting；
If horizontal distance of the 3D camera apart from human body is d, Human Height to be measured is h, the view of 3D camera vertical direction
Wild angle is θ, it is desirable that the scanning height of three 3D cameras is more than or equal to h, and formula is as follows:
d×6tan(θ/2)>h
D > h/6 × tan (θ/2),
Therefore, shot region is set as the border circular areas that diameter is not less than 2d；
It is H positioned at the height of nethermost camera from the ground1, calculate H1Formula are as follows: H1=tan (θ/2) × d；
It is H positioned at the height of uppermost camera from the ground3, calculate H3Formula are as follows: H3=h-tan (θ/2) × d；
Being located in the middle the height of camera from the ground is H2, calculate H2Formula are as follows: H2=(H1+H3)/2.
Further, after the installation 3D camera, further includes:
Using zhangShi standardization, respectively to built in 3D camera depth camera and RGB camera demarcate,
By Matlab Camera Calibration Toolbox program, the interior of depth camera and RGB camera is calculated separately out
Ginseng and outer ginseng matrix.
Further, the selection print area is respectively face, half body and whole body；
When selection face printing, system starts H3360 °/α platform 3D the camera of height；
When selecting half body printing, system starts H2、H32 × 360 °/α platform 3D camera of height；
When selecting whole body printing, system starts H1、H2、H33 × 360 °/α platform 3D camera of height.
Further, specific step is as follows for described image acquisition:
Firstly, depth camera measures the depth information in photographic subjects region, i.e. z coordinate, z coordinate represents photographic subjects
Depth away from depth camera visual angle；
RGB camera shoots coloured image target area, acquires the human body information of two-dimensional surface, i.e. x coordinate and y is sat
Mark, x coordinate represent the horizontal position of target object, and y-coordinate represents the height of target object.
Further, specific step is as follows for described image registration and three-dimensional point cloud generation:
Using RGB camera coordinate system as world coordinate system, then [Rrgb trgb]=[I 0], wherein R be camera rotation
Matrix, t are the translation vector of camera, if coordinate difference of the spatial point P under RGB camera coordinate system and depth camera coordinate system
For prgbAnd Pir, to formula Pir=Rir prgb+tirIt inverts, obtains prgb=Rir -1Pir-Rir -1tir, then spatial point P passes through (Rir -1-
Rir -1tir) the corresponding RGB coordinate of the available point of coordinate transform；
Rectangular coordinate system u, v as unit of being established respectively on RGB image by pixel and as unit of physical unit millimeter
Rectangular coordinate system XY, XY coordinate system using pixel O a certain in image as origin, two reference axis X, Y are parallel with u, v respectively；Its
In (u, v) indicate using pixel as the coordinate of the image coordinate system of basic unit, (X, Y) indicate using millimeter as the image of basic unit
The coordinate of coordinate system, the then coordinate points that two image coordinates are fastened have following coordinate conversion relation:
After depth camera coordinate system is transformed into RGB camera coordinate system, it is assumed that the depth coordinate of space any point p be (u,
V), then it is obtained by coordinate conversion relation:
In formula, xw、ywIt is spatial position of the point p in RGB camera coordinate system, zwIt is by depth camera coordinate system to RGB phase
The depth value of machine coordinate mooring points p, ɑxTo normalize focal length, ɑ on depth camera u axisyIt is burnt to be normalized on depth camera v axis
Away from (uo,vo) be depth camera picture centre coordinate, ɑx、ɑy、uo、voIt is related with the internal reference of camera；
According to above formula, the three-dimensional coordinate of point p is found out:
Aforesaid operations are all done to each of depth image pixel, obtain the three-dimensional coordinate of pixel on depth image from
And obtain the three-dimensional point cloud of body local.
Further, specific step is as follows for the multi-view image splicing:
Using SIFT operator, the characteristic point in each photo is found out, obtains exact position and the dimensional information of characteristic point, into
The matching of row characteristic point；
Complete characteristic point matching after obtain a Feature Points Matching set, selected from set of characteristic points minimum distance with
The secondary the smallest eight groups of matching characteristics point of short distance ratio carries out the matching of image, finds out the coordinate transform between two images；
The image of different perspectives is transformed into the same coordinate system, the initial registration of point cloud is completed, and uses stochastical sampling
Coherence method carries out geometric precision correction, obtains the Complete three-dimensional point cloud data of human body in photo.
Further, specific step is as follows for the building human 3d model:
Original Points Sample is sampled original point cloud data collection, is extracted key from two panels original point cloud using ICP algorithm
Point；
Initial corresponding point set estimation, two panels point cloud respectively to all key points of selection calculate separately its feature describe because
Son, binding characteristic describe the position of coordinate of the factor in two datasets, and the similarity with feature between the two and position is
Basis, to estimate their corresponding relationship, corresponding points pair according to a preliminary estimate, so that it is determined that initially corresponding to point set；
For mistake corresponding points to rejecting, the rejecting of mistake corresponding points pair uses the rejecting algorithm of stochastical sampling consistency；
Rigid body translation Matrix Solving is based on corresponding relationship estimation error module, estimates an initial transformation matrix, make
Transformation matrix is iterated to calculate with least square method, error function is minimized, until meeting convergence conditions；
It is registrated complete human body's point cloud data, threedimensional model construction terminates.
Compared with prior art, the invention has the following advantages that
The present invention provide it is a kind of print the integrated equipment of 3D portrait using self-service take pictures of 3D camera, equipment use more 3D
Camera, it is accurate to obtain human body color image and depth information data, by the way that the image characteristic point of different perspectives is merged, is registrated
And splicing, quickly create human 3d model, user can unrestricted choice print area, and pass through mobile payment, self-service completion 3D
Portrait printing.Integration degree is high for complete equipment, easy to operate, and visualized operation, measurement accuracy is higher, can be in unmanned value
The self-service whole process service for completing shooting, modeling, payment, printing in the case where keeping.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
It obtains other drawings based on these drawings.
Fig. 1 is that a kind of of the embodiment of the present invention prints the structure of the integrated equipment of 3D portrait using self-service take pictures of 3D camera
Block diagram；
Fig. 2 is that a kind of of the embodiment of the present invention prints the process signal of 3D portrait method using self-service take pictures of 3D camera
Figure；
Fig. 3 is one of the schematic diagram of 3D camera installation；
Fig. 4 is the two of the schematic diagram of 3D camera installation.
Meaning representated by serial number in figure: 11.3D camera, 12. data processing units, 13. display units, 14.3D are beaten
Impression member, 15. power supply units.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
The integrated equipment and method of 3D portrait are printed using self-service take pictures of 3D camera the embodiment of the invention provides a kind of,
The present invention is described in detail with specific embodiment below in conjunction with the accompanying drawings.
Embodiment one
It is provided in this embodiment it is a kind of print the integrated equipment of 3D portrait using self-service take pictures of 3D camera, which is divided into
Front shot region, middle part is shown and data processing region, rear portion print area；As shown in Figure 1, the equipment includes:
More 3D cameras 11, for shooting human body color image and obtaining human depth's information；
Display unit 13 is used for display operation order and image processing data；
Data processing unit 12, for the image information and depth information of the shooting of 3D camera described in calculation processing, and will
Image characteristic point fusion, registration and the splicing of different perspectives, quickly create human 3d model；
3D printing unit 14, for printing the 3D portrait of desired target area；
Power supply unit 15, for being more 3D cameras 11, display unit 13, data processing unit 12 and 3D printing unit
14 provide power supply.
The equipment will take pictures, the function of data processing and printing is collected on one device, and integration degree is high, and equipment is logical
The human body color image and depth information of 3D camera shooting corresponding region are crossed, and the image characteristic point of different perspectives is merged,
Registration and splicing quickly create human 3d model, and user uses mobile payment, and automatic printing simultaneously takes out 3D portrait, convenient fast
Victory, it is easy to operate.
The present embodiment also provide it is a kind of using the self-service method for printing 3D portrait of taking pictures of 3D camera, as shown in Fig. 2, comprising
Following steps:
Step S101 installs 3D camera
The horizontal view angle of current domestic 3D camera is 58~67 °, and vertical angle of view is 45~53 °, detecting distance 0.5~5
Rice, due to the limitation at visual angle, separate unit 3D camera single shot can only obtain the depth information and color image of body local, because
This needs more camera multi-angle of view photographic subjects regions to obtain whole depth information and color image, while guaranteeing to shoot
The 3D camera in region will have certain overlapping visual angle in vertical direction, in order to later data processing.
This equipment both horizontally and vertically installs more 3D cameras respectively, so that 3D camera coverage is covered
All spaces in lid shooting area, due to 3D camera between object at a distance from it is closer, shooting effect is more accurate, to reach most
Good shooting effect, this equipment shooting area are designed as circle.
It names a specific example and illustrates how arrangement 3D camera.
It 1, is 60 ° of estimations according to 3D camera horizontal direction field-of-view angle, as shown in figure 4, in the outer of round shot region
It encloses and is horizontally mounted 6 3D cameras (60 ° of interval), using 6 3D cameras from 6 directions to clap every 60 ° circle is located at
360 ° of Image Acquisition of horizontal direction are carried out according to the human body at regional center.
2,3 3D cameras are successively installed at interval of 60 ° of vertical direction in the periphery of round shot region.
3, horizontal distance of the 3D camera apart from human body is set as d, Human Height to be measured is h, it is desirable that 3 3D cameras are swept
It retouches height and is more than or equal to h, while ensuring that the vertical visual field of 3 3D cameras has overlapping region, overlapping region is 3 D human body number
According to the key that point cloud registering merges, the more more then later period point cloud registering fusions of overlapping are more accurate.Formula is as follows:
d×6tan(θ/2)>h
tan(θ/2)>h/d×6
d>h/6×tan(θ/2) (1)。
4, as shown in figure 3, being estimated according to minimum 45 ° of field-of-view angle θ of 3D camera vertical direction, it is assumed that user's height h
=2m is substituted into formula (1), and calculating d minimum should be 0.8 meter, and therefore, front shot region is the circle that diameter is not less than 1.6 meters
Region.
5, to ensure that nethermost 3D camera can scan the both feet of human body completely, terrain clearance can by geometrical relationship
: H1=tan (θ/2) × d, then can be calculated H1=0.33m, i.e., nethermost 3D camera 0.33m from the ground.
6, uppermost 3D camera terrain clearance is from the geometric relations: H3=h-tan (θ/2) × d, can be calculated H3
=1.67m, i.e., uppermost 3D camera 1.67m from the ground.
7, being located in the middle the height of 3D camera from the ground is H2, calculate H2Formula are as follows: H2=(H1+H3)/2 is calculated
H can be obtained2=1m, i.e., intermediate 3D camera 1m from the ground.
It is calculated according to this data, nethermost 3D camera can completely take the foot of human body, similarly, topmost
Camera can completely take the head of human body, it can be ensured that collect complete human body surface information.
Step S102, camera calibration
Using tool software, respectively to built in 3D camera depth camera and RGB camera demarcate.It uses
ZhangShi standardization respectively demarcates depth camera and RGB camera, passes through Matlab Camera
Calibration Toolbox program calculates the internal reference and outer ginseng matrix of depth camera and RGB camera.Image leader
Fixed purpose is that the point cloud in order to obtain the posture relationship between 3D camera, to obtain 3D camera is adjusted to same
Under coordinate system.Calibration process need to only execute once after the completion of system hardware platform building.
Step S103 selects print area
Before user takes pictures, first corresponding print area, respectively " face, half body, whole body " are selected on the display unit, three
Kind print area respectively corresponds the 3D camera of starting different number.
When selection face printing, system starts H36 3D cameras of height；
When selecting half body printing, system starts H2、H312 3D cameras of height；
When selecting whole body printing, system starts H1、H2、H318 3D cameras of height.
Step S104, selection payment
After selected print area, equipment display unit generates payment interface, and the difference (face of print area is selected according to user
Portion, half body, whole body), user pays the payment that two dimensional code completes printing expense by mobile scanning terminal.
Step S105, Image Acquisition
When shooting, equipment is stood by voice prompting user in the center of shot region, and is kept fixed posture 5 seconds
Clock, 3D camera start and work asynchronously, and carry out image data acquiring to human body with vertical direction from horizontal direction.
Firstly, the infrared transmitter of depth camera emits near infrared light, after the reflection of human body, infrared image is passed
Sensor receives the near infrared light returned, and deformation situation or the flight time of return light are calculated by processing chip, measures shooting mesh
The depth information in region is marked, i.e. z coordinate, z coordinate represents depth of the photographic subjects away from depth camera visual angle, i.e., take the photograph from depth
As the distance of head, coordinate unit is millimeter.
The RGB image sensor of RGB camera starts, and shoots coloured image to target area, acquires the people of two-dimensional surface
Body information, i.e. x coordinate and y-coordinate, x coordinate represent the horizontal position of target object, and y-coordinate represents the height of target object, and 3
The unit of coordinate is millimeter.
Depth information is mainly used for spatial coordinates calculation and generates three-dimensional point cloud, and colour information is mainly used for object pixel
The extraction of coordinate points and point cloud registering.3D camera transfers information to data processing unit, carries out data processing by computer.
The depth information that adjacent 3D camera obtains will guarantee that the lap for having certain, the registration of depth information are exactly
According to the overlapping region between two images, completely three-dimensional point cloud to be fitted could be obtained by the depth information of registration.
Step S106, image registration and three-dimensional point cloud generate
Position and visual angle difference due to the depth camera on 3D camera with RGB camera, need first to take the photograph depth
As head be coordinately transformed with realize the calibration of depth image and color image be registrated, then obtained by the information of depth image
The three-dimensional coordinate of color image finally encloses corresponding RGB color for each of depth image pixel to obtain people
The three-dimensional point cloud of body part.Specific step is as follows:
1, a world coordinate system is established to determine the relative position of video camera and object, and world coordinate system is by Xw、Yw、Zw
Axis composition.Since the coordinate of depth camera and the coordinate of RGB camera are different, can be contacted with a rotation translation transformation
Come.Using RGB camera coordinate system as world coordinate system, then [Rrgb trgb]=[I 0], wherein R be camera spin matrix,
T is the translation vector of camera, if spatial point P is under RGB camera coordinate system (world coordinate system) and depth camera coordinate system
Coordinate is respectively prgbAnd Pir, to formula Pir=Rir prgb+tirIt inverts, obtains prgb=Rir -1Pir-Rir -1tir, then spatial point P is passed through
Cross (Rir -1-Rir -1tir) the corresponding RGB coordinate of the available point of coordinate transform.
2, rectangular coordinate system u, v as unit of being established respectively on RGB image by pixel and be single with physical unit millimeter
For rectangular coordinate system XY, the XY coordinate system of position using pixel O a certain in image as origin, two reference axis X, Y are parallel with u, v respectively；
Wherein (u, v) indicates that, using pixel as the coordinate of the image coordinate system of basic unit, (X, Y) is indicated using millimeter as the figure of basic unit
As the coordinate of coordinate system, then the coordinate points that two image coordinates are fastened have following coordinate conversion relation:
3, after depth camera coordinate system is transformed into RGB camera coordinate system, it is assumed that the depth coordinate of space any point p is
(u, v) is then obtained by coordinate conversion relation:
In formula, xw、ywIt is spatial position of the point p in RGB camera coordinate system, zwIt is by depth camera coordinate system to RGB phase
The depth value of machine coordinate mooring points p, ɑxTo normalize focal length, ɑ on depth camera u axisyIt is burnt to be normalized on depth camera v axis
Away from (uo,vo) be depth camera picture centre coordinate, ɑx、ɑy、uo、voIt is related with the internal reference of camera；
According to above formula, the three-dimensional coordinate of point p is found out:
The coordinate of point p is a three-dimensional vector, and x and y coordinates are that pixel of this in RGB camera coordinate system is sat
Mark, takes out the color of the pixel, the color as respective pixel in depth image.According to formula Pir=Rir prgb+tir, to depth
Each of degree image pixel all does aforesaid operations, obtains the three-dimensional coordinate of pixel on depth image to obtain human body office
The three-dimensional point cloud in portion.
Step S107, multi-view image splicing
Body local point cloud data after step S106 image registration, each point are stored with 3 floating-point values, and 3
Numerical value respectively represents the three dimensional space coordinate information of human body, i.e. x coordinate, the numerical value of y-coordinate and z coordinate, while each cloud is also
3 bytes are stored with, the colouring information indicated using RGB is represented, 3 bytes respectively correspond the value of R, G, B.Due to every 3D
There is lap at camera visual angle, need to splice and be registrated the body local point cloud that depth image represents under different perspectives, can obtain
Obtain complete 3 D human body point cloud data.Specific step is as follows:
1, using SIFT operator, the characteristic point in each photo is found out, obtains exact position and the dimensional information of characteristic point,
Carry out the matching of characteristic point；
2, a Feature Points Matching set is obtained after completing the matching of characteristic point, is then selected from set of characteristic points recently
Distance carries out the matching of image with the smallest eight groups of matching characteristics point of secondary short distance ratio, and the coordinate found out between two images becomes
It changes；
3, the image of different perspectives is transformed into the same coordinate system, completes the initial registration of point cloud, and use and adopt at random
Sample consistency (RANSAC) method carries out geometric precision correction, obtains the Complete three-dimensional point cloud data of human body in photo.
Step S108 constructs human 3d model
According to human body three-dimensional point cloud data, data processing unit is based on human body using the method for the dense matchings such as ICP algorithm
Point cloud carries out Poisson resurfacing and obtains human 3d model, ultimately produces corresponding data file and saves.Specific step is as follows:
1, original Points Sample is sampled original point cloud data collection, is extracted pass from two panels original point cloud using ICP algorithm
Key point；
2, initially corresponding point set estimation, two panels point cloud respectively to all key points of selection calculate separately its feature describe because
Son, binding characteristic describe the position of coordinate of the factor in two datasets, and the similarity with feature between the two and position is
Basis, to estimate their corresponding relationship, corresponding points pair according to a preliminary estimate, so that it is determined that initially corresponding to point set；
3, mistake corresponding points are to rejecting.Initial corresponding points concentrate not every corresponding relationship all correct, the correspondence of mistake
Point reduces point cloud registering accuracy to the estimated result that will have a direct impact on rigid body translation matrix.Mistake corresponding points are to picking
Except stochastical sampling consistency (RANSAC) or other rejecting algorithms can be used；In order to improve the estimated accuracy of transformation matrix and match
Quasi velosity, ICP algorithm only use a subset for removing initial corresponding point set of the wrong corresponding points to after as algorithm iteration
Initial point set；
4, rigid body translation Matrix Solving is based on corresponding relationship estimation error module, estimates an initial transformation matrix,
Transformation matrix is iterated to calculate using least square method, error function is minimized, until meeting convergence conditions；
5, it is registrated complete human body's point cloud data, threedimensional model construction terminates.
Step S109 prints 3D portrait
The human body 3D image data in region to be printed is transferred to the 3D printing unit at equipment rear portion by data processing unit, is beaten
Print machine starts to print 3D portrait.
Step S110 takes out 3D portrait
After completing printing, user clicks completing button in display unit, front shot region and rear portion print area every
Plate cancels locking state, and user can voluntarily take out 3D portrait, and whole printing process finishes.
Finally, it should be noted that embodiment described above, only a specific embodiment of the invention, to illustrate the present invention
Technical solution, rather than its limitations, scope of protection of the present invention is not limited thereto, although with reference to the foregoing embodiments to this hair
It is bright to be described in detail, those skilled in the art should understand that: anyone skilled in the art
In the technical scope disclosed by the present invention, it can still modify to technical solution documented by previous embodiment or can be light
It is readily conceivable that variation or equivalent replacement of some of the technical features；And these modifications, variation or replacement, do not make
The essence of corresponding technical solution is detached from the spirit and scope of technical solution of the embodiment of the present invention, should all cover in protection of the invention
Within the scope of.Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (9)
1. a kind of print the integrated equipment of 3D portrait using self-service take pictures of 3D camera, which is characterized in that the equipment includes:
More 3D cameras, for shooting human body color image and obtaining human depth's information；
Display unit is used for display operation order and image processing data；
Data processing unit, for the image information and depth information of the shooting of 3D camera described in calculation processing, and by different views
Image characteristic point fusion, registration and the splicing at angle, quickly create human 3d model；
3D printing unit, for printing the 3D portrait of desired target area；
Power supply unit, for providing power supply for more 3D cameras, display unit, data processing unit and 3D printing unit.
2. it is a kind of using it is described in claim 1 using 3D camera it is self-service take pictures print 3D portrait integrated equipment realize from
Help the method taken pictures and print 3D portrait, which is characterized in that comprise the steps of:
3D camera is installed, more 3D cameras are both horizontally and vertically installed respectively, so that 3D camera shoots model
Enclose all spaces in covering shooting area；
Print area is selected, before user takes pictures, first selects corresponding print area on the display unit, different print areas are corresponding
Start the 3D camera of different number；
Selection payment, after selecting print area, the display unit of equipment generates payment interface, and user passes through mobile scanning terminal branch
Pay the payment that two dimensional code completes printing expense；
Image Acquisition, 3D camera start work, carry out image data acquiring, image to human body with vertical direction from horizontal direction
Data include the color image and depth information of human body；
Image registration and three-dimensional point cloud generate, and enclose corresponding RGB color for each of depth image pixel to obtain
To the three-dimensional point cloud of body local；
Multi-view image splicing, since there is lap at every 3D camera visual angle, need to splice and be registrated depth under different perspectives
The body local point cloud that image represents, can be obtained complete 3 D human body point cloud data；
Human 3d model is constructed, Poisson resurfacing is carried out based on human body point cloud and obtains human 3d model；
3D portrait is printed, the 3D image data of region human body to be printed is transferred to 3D printing unit printing 3D by data processing unit
Portrait.
3. according to claim 2 use the self-service method for printing 3D portrait of taking pictures of 3D camera, which is characterized in that described
Installing 3D camera, specific step is as follows:
Due to 3D camera between object at a distance from it is closer, shooting effect is more accurate, to reach best shooting effect, shot region
It is designed as circle, more 3D cameras are both horizontally and vertically installed respectively；
If 3D camera horizontal direction field-of-view angle is α, 360 °/α platform 3D camera shooting is horizontally mounted in the periphery of round shot region
Head, 360 °/α platform 3D camera is to carry out 360 ° of images of horizontal direction to the human body being located at round shot region center every α
Acquisition；
In round shot region at interval of three 3D cameras of α right angle setting；
If horizontal distance of the 3D camera apart from human body is d, Human Height to be measured is h, the angle of visibility of 3D camera vertical direction
Degree is θ, it is desirable that the scanning height of three 3D cameras is more than or equal to h, and formula is as follows:
d×6tan(θ/2)>h
D > h/6 × tan (θ/2),
Therefore, shot region is set as the border circular areas that diameter is not less than 2d；
It is H positioned at the height of nethermost camera from the ground1, calculate H1Formula are as follows: H1=tan (θ/2) × d；
It is H positioned at the height of uppermost camera from the ground3, calculate H3Formula are as follows: H3=h-tan (θ/2) × d；
Being located in the middle the height of camera from the ground is H2, calculate H2Formula are as follows: H2=(H1+H3)/2.
4. according to claim 2 use the self-service method for printing 3D portrait of taking pictures of 3D camera, which is characterized in that described
After installation 3D camera, further includes:
Using zhangShi standardization, respectively to built in 3D camera depth camera and RGB camera demarcate, pass through
Matlab Camera Calibration Toolbox program, calculate separately out depth camera and RGB camera internal reference and
Outer ginseng matrix.
5. according to claim 3 use the self-service method for printing 3D portrait of taking pictures of 3D camera, which is characterized in that described
Selecting print area is respectively face, half body and whole body；
When selection face printing, system starts H3360 °/α platform 3D the camera of height；
When selecting half body printing, system starts H2、H32 × 360 °/α platform 3D camera of height；
When selecting whole body printing, system starts H1、H2、H33 × 360 °/α platform 3D camera of height.
6. according to claim 4 use the self-service method for printing 3D portrait of taking pictures of 3D camera, which is characterized in that described
Specific step is as follows for Image Acquisition:
Firstly, depth camera measures the depth information in photographic subjects region, i.e. z coordinate, z coordinate represents photographic subjects away from depth
Spend the depth at camera visual angle；
RGB camera shoots coloured image target area, acquires the human body information of two-dimensional surface, i.e. x coordinate and y-coordinate, and x is sat
Mark represents the horizontal position of target object, and y-coordinate represents the height of target object.
7. according to claim 6 use the self-service method for printing 3D portrait of taking pictures of 3D camera, which is characterized in that described
Specific step is as follows for image registration and three-dimensional point cloud generation:
Using RGB camera coordinate system as world coordinate system, then [Rrgb trgb]=[I 0], wherein R be camera spin moment
Battle array, t is the translation vector of camera, if coordinate of the spatial point P under RGB camera coordinate system and depth camera coordinate system is respectively
prgbAnd Pir, to formula Pir=Rirprgb+tirIt inverts, obtains prgb=Rir -1Pir-Rir -1tir, then spatial point P passes through (Rir -1-Rir -1tir) the corresponding RGB coordinate of the available point of coordinate transform；
Rectangular coordinate system u, v as unit of being established respectively on RGB image by pixel and straight as unit of physical unit millimeter
For angular coordinate system XY, XY coordinate system using pixel O a certain in image as origin, two reference axis X, Y are parallel with u, v respectively；Wherein (u,
V) indicate that, using pixel as the coordinate of the image coordinate system of basic unit, (X, Y) is indicated using millimeter as the image coordinate of basic unit
The coordinate of system, the then coordinate points that two image coordinates are fastened have following coordinate conversion relation:
After depth camera coordinate system is transformed into RGB camera coordinate system, it is assumed that the depth coordinate of space any point p is (u, v), then
It is obtained by coordinate conversion relation:
In formula, xw、ywIt is spatial position of the point p in RGB camera coordinate system, zwIt is to be sat by depth camera coordinate system to RGB camera
Mark the depth value of mooring points p, ɑxTo normalize focal length, ɑ on depth camera u axisyTo normalize focal length on depth camera v axis,
(uo,vo) be depth camera picture centre coordinate, ɑx、ɑy、uo、voIt is related with the internal reference of camera；
According to above formula, the three-dimensional coordinate of point p is found out:
Aforesaid operations are all done to each of depth image pixel, obtain the three-dimensional coordinate of pixel on depth image to obtain
To the three-dimensional point cloud of body local.
8. according to claim 2 use the self-service method for printing 3D portrait of taking pictures of 3D camera, which is characterized in that described
Specific step is as follows for multi-view image splicing:
Using SIFT operator, the characteristic point in each photo is found out, obtains exact position and the dimensional information of characteristic point, is carried out special
Levy the matching of point；
A Feature Points Matching set is obtained after completing the matching of characteristic point, and minimum distance and time close is selected from set of characteristic points
The smallest eight groups of matching characteristics point of distance ratio carries out the matching of image, finds out the coordinate transform between two images；
The image of different perspectives is transformed into the same coordinate system, completes the initial registration of point cloud, and consistent using stochastical sampling
Property method carry out geometric precision correction, obtain the Complete three-dimensional point cloud data of human body in photo.
9. according to claim 8 use the self-service method for printing 3D portrait of taking pictures of 3D camera, which is characterized in that described
Constructing human 3d model, specific step is as follows:
Original Points Sample samples original point cloud data collection, extracts key point from two panels original point cloud using ICP algorithm；
Initial corresponding point set estimation, two panels point cloud respectively calculate separately its feature to all key points of selection and describe the factor, tie
The position that feature describes coordinate of the factor in two datasets is closed, based on the similarity of feature between the two and position,
Their corresponding relationship is estimated, according to a preliminary estimate corresponding points pair, so that it is determined that initially corresponding to point set；
For mistake corresponding points to rejecting, the rejecting of mistake corresponding points pair uses the rejecting algorithm of stochastical sampling consistency；
Rigid body translation Matrix Solving is based on corresponding relationship estimation error module, an initial transformation matrix is estimated, using most
Small square law iterates to calculate transformation matrix, error function is minimized, until meeting convergence conditions；
It is registrated complete human body's point cloud data, threedimensional model construction terminates.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811058666.7A CN109304866A (en) | 2018-09-11 | 2018-09-11 | The integrated equipment and method of 3D portrait are printed using self-service take pictures of 3D camera |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811058666.7A CN109304866A (en) | 2018-09-11 | 2018-09-11 | The integrated equipment and method of 3D portrait are printed using self-service take pictures of 3D camera |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109304866A true CN109304866A (en) | 2019-02-05 |
Family
ID=65224933
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811058666.7A Pending CN109304866A (en) | 2018-09-11 | 2018-09-11 | The integrated equipment and method of 3D portrait are printed using self-service take pictures of 3D camera |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109304866A (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109823070A (en) * | 2019-04-01 | 2019-05-31 | 广州四两科技有限公司 | A kind of Intelligent printing method of clothes |
| CN111150175A (en) * | 2019-12-05 | 2020-05-15 | 新拓三维技术(深圳)有限公司 | Method, device and system for three-dimensional scanning of feet |
| CN111199579A (en) * | 2020-01-02 | 2020-05-26 | 腾讯科技（深圳）有限公司 | Method, device, equipment and medium for building three-dimensional model of target object |
| CN111445559A (en) * | 2020-04-07 | 2020-07-24 | 珠海格力电器股份有限公司 | Image processing and displaying method, device, equipment and computer readable medium |
| CN111556255A (en) * | 2020-04-30 | 2020-08-18 | 华为技术有限公司 | Image generation method and device |
| CN111932599A (en) * | 2020-07-31 | 2020-11-13 | 泉州市澳莱格电子有限责任公司 | Cylinder two-dimensional image generation method based on multiple RGB-D cameras |
| CN112118438A (en) * | 2020-06-30 | 2020-12-22 | 中兴通讯股份有限公司 | Camera system, mobile terminal and three-dimensional image acquisition method |
| CN112146564A (en) * | 2019-06-28 | 2020-12-29 | 先临三维科技股份有限公司 | Three-dimensional scanning method, three-dimensional scanning device, computer equipment and computer readable storage medium |
| CN113240603A (en) * | 2020-12-28 | 2021-08-10 | 武汉纺织大学 | Three-dimensional human body measuring system and method based on point cloud data |
| CN113256721A (en) * | 2021-06-21 | 2021-08-13 | 浙江光珀智能科技有限公司 | Indoor multi-person three-dimensional high-precision positioning method |
| CN113823044A (en) * | 2021-10-08 | 2021-12-21 | 刘智矫 | Human body three-dimensional data acquisition room and charging method thereof |
| CN113963062A (en) * | 2021-11-03 | 2022-01-21 | 魏帅 | Camera array positioning method of virtual fitting system |
| CN116080290A (en) * | 2022-12-29 | 2023-05-09 | 上海魅奈儿科技有限公司 | Three-dimensional high-precision fixed-point printing method and device |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104275944A (en) * | 2013-07-08 | 2015-01-14 | 西安非凡士机器人科技有限公司 | Rapid manufacturing system for human body three-dimensional model |
| CN207755278U (en) * | 2017-06-27 | 2018-08-24 | 西安蒜泥电子科技有限责任公司 | A kind of human body three-dimensional scanning system |
-
2018
- 2018-09-11 CN CN201811058666.7A patent/CN109304866A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104275944A (en) * | 2013-07-08 | 2015-01-14 | 西安非凡士机器人科技有限公司 | Rapid manufacturing system for human body three-dimensional model |
| CN207755278U (en) * | 2017-06-27 | 2018-08-24 | 西安蒜泥电子科技有限责任公司 | A kind of human body three-dimensional scanning system |
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109823070A (en) * | 2019-04-01 | 2019-05-31 | 广州四两科技有限公司 | A kind of Intelligent printing method of clothes |
| CN112146564B (en) * | 2019-06-28 | 2022-04-15 | 先临三维科技股份有限公司 | Three-dimensional scanning method, three-dimensional scanning device, computer equipment and computer readable storage medium |
| CN112146564A (en) * | 2019-06-28 | 2020-12-29 | 先临三维科技股份有限公司 | Three-dimensional scanning method, three-dimensional scanning device, computer equipment and computer readable storage medium |
| CN111150175A (en) * | 2019-12-05 | 2020-05-15 | 新拓三维技术(深圳)有限公司 | Method, device and system for three-dimensional scanning of feet |
| CN111199579A (en) * | 2020-01-02 | 2020-05-26 | 腾讯科技（深圳）有限公司 | Method, device, equipment and medium for building three-dimensional model of target object |
| CN111199579B (en) * | 2020-01-02 | 2023-01-24 | 腾讯科技（深圳）有限公司 | Method, device, equipment and medium for building three-dimensional model of target object |
| WO2021135627A1 (en) * | 2020-01-02 | 2021-07-08 | 腾讯科技（深圳）有限公司 | Method for constructing three-dimensional model of target object, and related apparatus |
| US12014461B2 (en) | 2020-01-02 | 2024-06-18 | Tencent Technology (Shenzhen) Company Limited | Method for constructing three-dimensional model of target object and related apparatus |
| CN111445559A (en) * | 2020-04-07 | 2020-07-24 | 珠海格力电器股份有限公司 | Image processing and displaying method, device, equipment and computer readable medium |
| CN111556255B (en) * | 2020-04-30 | 2021-10-01 | 华为技术有限公司 | Image generation method and device |
| CN111556255A (en) * | 2020-04-30 | 2020-08-18 | 华为技术有限公司 | Image generation method and device |
| CN112118438A (en) * | 2020-06-30 | 2020-12-22 | 中兴通讯股份有限公司 | Camera system, mobile terminal and three-dimensional image acquisition method |
| CN111932599A (en) * | 2020-07-31 | 2020-11-13 | 泉州市澳莱格电子有限责任公司 | Cylinder two-dimensional image generation method based on multiple RGB-D cameras |
| CN113240603A (en) * | 2020-12-28 | 2021-08-10 | 武汉纺织大学 | Three-dimensional human body measuring system and method based on point cloud data |
| CN113256721A (en) * | 2021-06-21 | 2021-08-13 | 浙江光珀智能科技有限公司 | Indoor multi-person three-dimensional high-precision positioning method |
| CN113823044B (en) * | 2021-10-08 | 2022-09-13 | 刘智矫 | Human body three-dimensional data acquisition room and charging method thereof |
| CN113823044A (en) * | 2021-10-08 | 2021-12-21 | 刘智矫 | Human body three-dimensional data acquisition room and charging method thereof |
| CN113963062A (en) * | 2021-11-03 | 2022-01-21 | 魏帅 | Camera array positioning method of virtual fitting system |
| CN116080290A (en) * | 2022-12-29 | 2023-05-09 | 上海魅奈儿科技有限公司 | Three-dimensional high-precision fixed-point printing method and device |
| WO2024140186A1 (en) * | 2022-12-29 | 2024-07-04 | 上海魅奈儿科技有限公司 | Three-dimensional high-precision positioned printing method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109304866A (en) | The integrated equipment and method of 3D portrait are printed using self-service take pictures of 3D camera | |
| CN108470370B (en) | Method for jointly acquiring three-dimensional color point cloud by external camera of three-dimensional laser scanner | |
| CN108151671B (en) | A kind of 3 D digital imaging sensor, 3 D scanning system and its scan method | |
| CN104335005B (en) | 3D is scanned and alignment system | |
| CN106289106B (en) | The stereo vision sensor and scaling method that a kind of line-scan digital camera and area array cameras are combined | |
| Bonfort et al. | Voxel carving for specular surfaces | |
| AU2011312140B2 (en) | Rapid 3D modeling | |
| JP4245963B2 (en) | Method and system for calibrating multiple cameras using a calibration object | |
| US7751651B2 (en) | Processing architecture for automatic image registration | |
| CN103559735B (en) | A kind of three-dimensional rebuilding method and system | |
| Bonfort et al. | General specular surface triangulation | |
| CN104657103B (en) | Hand-held CAVE optical projection systems based on depth camera | |
| CN106780618A (en) | 3 D information obtaining method and its device based on isomery depth camera | |
| Shashi et al. | Use of photogrammetry in 3D modeling and visualization of buildings | |
| JP2003130621A (en) | Method and system for measuring three-dimensional shape | |
| CN109357633A (en) | 3-D scanning method, apparatus, storage medium and processor | |
| JP2005077385A (en) | Image correlation method, survey method and measuring system using them | |
| CN107610219A (en) | The thick densification method of Pixel-level point cloud that geometry clue perceives in a kind of three-dimensional scenic reconstruct | |
| Reitinger et al. | Augmented reality scouting for interactive 3d reconstruction | |
| CN108257182A (en) | A kind of scaling method and device of three-dimensional camera module | |
| CN108175535A (en) | A kind of dentistry spatial digitizer based on microlens array | |
| CN111340959A (en) | Three-dimensional model seamless texture mapping method based on histogram matching | |
| CN101448088B (en) | Initial location system in virtual studio | |
| Frisky et al. | Acquisition Evaluation on Outdoor Scanning for Archaeological Artifact Digitalization. | |
| CN110288654A (en) | A kind of method that the geometry of single image measures |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190205 |