CN106931969A - A kind of robot three-dimensional navigation map generation method based on Kinect - Google Patents
A kind of robot three-dimensional navigation map generation method based on Kinect Download PDFInfo
- Publication number
- CN106931969A CN106931969A CN201511009650.3A CN201511009650A CN106931969A CN 106931969 A CN106931969 A CN 106931969A CN 201511009650 A CN201511009650 A CN 201511009650A CN 106931969 A CN106931969 A CN 106931969A
- Authority
- CN
- China
- Prior art keywords
- kinect
- robot
- face
- method based
- navigation map
- 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
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/20—Instruments for performing navigational calculations
- G01C21/206—Instruments for performing navigational calculations specially adapted for indoor navigation
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Automation & Control Theory (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Processing Or Creating Images (AREA)
- Image Processing (AREA)
Abstract
A kind of robot three-dimensional navigation map generation method based on Kinect, gathers Kinect initial data；Laplce's smoothing processing, the preliminary alignment in face；Based on the resurfacing of segmentation, the texture mapping based on visual angle.The present invention instructs Indoor Robot to be moved for the path planning and avoidance of indoor mobile robot, cheap with equipment price, operates easy advantage, and suitable domestic consumer uses.
Description
Technical field
The present invention relates to a kind of path planning for indoor mobile robot and avoidance technology, a kind of three-dimensional is related specifically to
Navigation map generation method.
Background technology
Map building technology has a very important role for the path planning of mobile robot, first to two dimension in robot chamber
Ground drawing generating method is described, including Grid Method, asdic method, Stereo Vision etc..With the development of science and technology,
Three-dimensional map creates technology turns into study hotspot.Kinect is a kind of feeling device, possesses more powerful image processing function.Profit
Indoor environment is scanned with Kinect, after the data that will be obtained are processed through computer, the three-dimensional map of indoor environment is obtained.
The content of the invention
It is an object of the invention to provide a kind of indoor mobile robot three-dimensional navigation ground drawing generating method, indoor machine is effectively instructed
People moves.
A kind of robot three-dimensional navigation map generation method based on Kinect, it is characterised in that：It is original by gathering Kinect
Data；Carry out Laplce's smoothing processing；The preliminary alignment in face；Resurfacing based on segmentation；Texture mapping based on visual angle.
Described a kind of robot three-dimensional navigation map generation method based on Kinect, it is characterised in that described Kinect is original
Collecting method is：The data that Kinect is collected are filled to buffering area first, is then added to depth data therein
Accumulation buffer, is then added with the data of former buffering area and averages, if this average value is certain less than what is be previously set
Threshold value, then it is just cumulative, can not otherwise be added up, accumulative frequency is set to 5 here.
Described a kind of robot three-dimensional navigation map generation method based on Kinect, it is characterized in that described Laplce smooths
Processing method is：By the three-dimensional situation of the surrounding vertex of umbrella apposition, offset to its position of centre of gravity.Then for each on face
Individual summit, relation and point according to this point and surrounding point calculate the sky of this point in theory again in the positional information in space
Between position.Shown in smoothing processing formula such as formula (1)：
In above-mentioned formulaIt is the position that i-th summit is newest, N is to want to ask the summit around the point of locus again
Number, can determine the scope of N according to experiment.
Described a kind of robot three-dimensional navigation map generation method based on Kinect, it is characterized in that the preliminary alignment in described face
Method is：Using surface in alignment-semi-automatic method：First determine a datum level, then determine the determination by former and later two faces substantially
The profile of object.
Described a kind of robot three-dimensional navigation map generation method based on Kinect, it is characterized in that the described table based on segmentation
Face method for reconstructing is：A cloud generation Octree is first passed through, the outer surface data based on these clouds is then obtained, from a cloud to face
The data conversion that Kinect can be collected of reconstruction into conventional surface data, so obtain a several independent faces for object,
These independent faces are merged again, the outer surface profile of a cloud can be obtained after merging.
Described a kind of robot three-dimensional navigation map generation method based on Kinect, it is characterized in that the described line based on visual angle
Managing chart pasting method is：View directions are unit vector of the required surface point to camera position, and face n-tuple is the top of body surface
To the unit vector of virtual Kinect positions, when particular curvature is shot, virtual Kinect directions are current point and Kinect to point
Direction.Wn is the angle cosine value of view directions and face n directions.When final texture color is calculated, weighted average is taken,
So that the color on each summit is together decided on by each surface texture.
The essence of technical scheme is：Raw data acquisition is carried out first with Kinect, environment depth information is obtained
Data, then carry out Laplce's smoothing processing, to reduce the influence of noise, by the preliminary of face to the view data for collecting
After alignment, precise transformation matrix is obtained using iteration closest approach algorithm, to obtain coordinate streamline；Then carried out using split plot design
Resurfacing and texture mapping, so that model is truer, finally derive threedimensional model.
The advantage of the technical scheme is：
Equipment cost is low, and operation is easy, and suitable domestic consumer uses.
Brief description of the drawings
Fig. 1 is that earth-magnetism navigation reference map builds flow chart；
Fig. 2 indoor environment plans；
Specific embodiment
Step 1：Collection Kinect initial data：
The data that Kinect is collected are filled to buffering area first, depth data therein are then added to accumulation buffer,
Then it is added with the data of former buffering area and is averaged, if this average value is less than the certain threshold value being previously set, then
It is just cumulative, can not otherwise be added up, accumulative frequency is set to 5 here.
Step 2：Laplce's smoothing processing：
By the three-dimensional situation of the surrounding vertex of umbrella apposition, offset to its position of centre of gravity.Then for each summit on face,
Relation and point according to this point and surrounding point calculate the locus of this point in theory again in the positional information in space.
Shown in smoothing processing formula such as formula (1)：
In above-mentioned formulaIt is the position that i-th summit is newest, N is to want to ask the summit around the point of locus again
Number.
The value of N can not it is excessive can not be too small, it is excessive it is too small can all bring certain influence, and this influence is passive.
If N values are too small, then cannot just ensure that the smooth effect for obtaining meets actual requirement, but N values are too greatly,
Can cause that treatment of details is undesirable again, therefore the selection of N values is that have certain scope, and expected effect can be just reached within this range
Really, the scope of N can determine according to experiment.
Step 3：The preliminary alignment in face：
Using surface in alignment-semi-automatic method：First determine a datum level, then determine by former and later two faces substantially earnest body really
Profile.
After computation the face in face relative to datum level transformation matrix when, bounding box center below is translated first, translate
To the center of the rear side rectangle of the bounding box of datum level.180 degree will be rotated according to z-axis below again, finally carry out curved surface profile
Extract.It is that the three-dimensional point range obtained after step before is projected into X/Y plane that curved surface profile extracts specific practice, obtains one
Plane mask figure.After the extraction of front and rear two curved surface profile, constantly rotation translates curved surface below, makes the wheel of its profile and front curve
Wide closest, the calculating of distance herein uses the distance between point cloud under Cartesian coordinates.
In the projected outline of YZ planes calculates, the value of virtual Kinect positions is left surface in data acquisition to front and rear composite surface
During Kinect position.The computational methods of the virtual Kinect positions in right side are similar.Obtain virtual Kinect positions it
Afterwards, by reverse projection matrix, the two dimensional image that three-dimensional body is projected in perspective plane can be calculated, by this two dimension
The method of the calculating projection that image can be used before and after in the alignment of two sides, calculates the projection on left and right two sides.
Step 4：Resurfacing based on segmentation：
A cloud generation Octree is first passed through, the outer surface data based on these clouds are then obtained, can to the reconstruction in face from a cloud
The data conversion that Kinect is collected so obtains a several independent faces for object into conventional surface data, then this is several
Individual independent face merges, and the outer surface profile of a cloud can be obtained after merging.
Step 5：Texture mapping method based on visual angle is：
View directions are unit vector of the required surface point to camera position, and face n-tuple is the summit of body surface to virtually
The unit vector of Kinect positions, when particular curvature is shot, virtual Kinect directions are the direction of current point and Kinect.
Wn is the angle cosine value of view directions and face n directions.When final texture color is calculated, weighted average is taken so that every
One color on summit is together decided on by each surface texture.
Claims (6)
1. a kind of robot three-dimensional navigation map generation method based on Kinect, it is characterised in that：It is former by gathering Kinect
Beginning data；Carry out Laplce's smoothing processing；The preliminary alignment in face；Resurfacing based on segmentation；Texture patch based on visual angle
Figure.
2. a kind of robot three-dimensional navigation map generation method based on Kinect according to claim 1, its feature exists
It is in described Kinect raw data acquisition methods：The data that Kinect is collected are filled to buffering area, then by it first
In depth data be added to accumulation buffer, be then added with the data of former buffering area and averaged, if this average value
Less than the certain threshold value being previously set, then just cumulative, can not otherwise be added up, accumulative frequency is set to 5 here.
3. a kind of robot three-dimensional navigation map generation method based on Kinect according to claim 1, it is characterized in that
Described Laplce's smoothing processing method is：By the three-dimensional situation of the surrounding vertex of umbrella apposition, offset to its position of centre of gravity.
Then for each summit on face, relation and point according to this point and surrounding point exist again in the positional information in space
The locus of this point is calculated in theory.Shown in smoothing processing formula such as formula (1)：
In above-mentioned formulaIt is the position that i-th summit is newest, N is to want to ask the summit around the point of locus again
Number, can determine the scope of N according to experiment.
4. a kind of robot three-dimensional navigation map generation method based on Kinect according to claim 1, it is characterized in that
The preliminary alignment schemes in described face are：Using surface in alignment-semi-automatic method：First determine a datum level, then determine front and rear two
Individual face can the substantially profile of earnest body really.
5. a kind of robot three-dimensional navigation map generation method based on Kinect according to claim 1, it is characterized in that
The described method of surface reconstruction based on segmentation is：A cloud generation Octree is first passed through, the appearance based on these clouds is then obtained
Face data, the data conversion that the reconstruction from a cloud to face can collect Kinect so obtains one into conventional surface data
Several independent faces of individual object, then these independent faces are merged, the outer surface profile of a cloud can be obtained after merging.
6. a kind of robot three-dimensional navigation map generation method based on Kinect according to claim 1, it is characterized in that
The described texture mapping method based on visual angle is：View directions are unit vector of the required surface point to camera position, face n
Vector is the summit of body surface to the unit vector of virtual Kinect positions, when particular curvature is shot, virtual Kinect directions
It is current point and the direction of Kinect.Wn is the angle cosine value of view directions and face n directions.Calculating final texture color
When, take weighted average so that the color on each summit is together decided on by each surface texture.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201511009650.3A CN106931969A (en) | 2015-12-29 | 2015-12-29 | A kind of robot three-dimensional navigation map generation method based on Kinect |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201511009650.3A CN106931969A (en) | 2015-12-29 | 2015-12-29 | A kind of robot three-dimensional navigation map generation method based on Kinect |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106931969A true CN106931969A (en) | 2017-07-07 |
Family
ID=59457538
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201511009650.3A Pending CN106931969A (en) | 2015-12-29 | 2015-12-29 | A kind of robot three-dimensional navigation map generation method based on Kinect |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106931969A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107844292A (en) * | 2017-10-31 | 2018-03-27 | 苏州乐米信息科技股份有限公司 | The control method and its system of camera curves track in a kind of cool run game |
| CN111216124A (en) * | 2019-12-02 | 2020-06-02 | 广东技术师范大学 | Robot vision guiding method and device based on integration of global vision and local vision |
-
2015
- 2015-12-29 CN CN201511009650.3A patent/CN106931969A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107844292A (en) * | 2017-10-31 | 2018-03-27 | 苏州乐米信息科技股份有限公司 | The control method and its system of camera curves track in a kind of cool run game |
| CN111216124A (en) * | 2019-12-02 | 2020-06-02 | 广东技术师范大学 | Robot vision guiding method and device based on integration of global vision and local vision |
| CN111216124B (en) * | 2019-12-02 | 2020-11-06 | 广东技术师范大学 | Robot vision guiding method and device based on integration of global vision and local vision |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104677347A (en) | Indoor mobile robot capable of producing 3D navigation map based on Kinect | |
| CN108520554B (en) | Binocular three-dimensional dense mapping method based on ORB-SLAM2 | |
| CN102136155B (en) | Object elevation vectorization method and system based on three dimensional laser scanning | |
| Concha et al. | Using superpixels in monocular SLAM | |
| CN104376596B (en) | A kind of three-dimensional scene structure modeling and register method based on single image | |
| CN102880866B (en) | Method for extracting face features | |
| CN100559398C (en) | Automatic deepness image registration method | |
| CN104966316A (en) | 3D face reconstruction method, apparatus and server | |
| CN111508074A (en) | Three-dimensional building model simplification method based on roof contour line | |
| US20130124148A1 (en) | System and Method for Generating Editable Constraints for Image-based Models | |
| CN104616286B (en) | Quick semi-automatic multi views depth restorative procedure | |
| CN105354883A (en) | 3ds Max fast and precise three-dimensional modeling method and system based on point cloud | |
| CN104661010A (en) | Method and device for establishing three-dimensional model | |
| CN109242954A (en) | Multi-view angle three-dimensional human body reconstruction method based on template deformation | |
| CN107730587B (en) | Rapid three-dimensional interactive modeling method based on pictures | |
| CN103854301A (en) | 3D reconstruction method of visible shell in complex background | |
| CN102081733B (en) | Multi-modal information combined pose-varied three-dimensional human face five-sense organ marking point positioning method | |
| CN108564619B (en) | Realistic three-dimensional face reconstruction method based on two photos | |
| CN107545602B (en) | Building modeling method under space topological relation constraint based on LiDAR point cloud | |
| CN115564926A (en) | Three-dimensional patch model construction method based on image building structure learning | |
| CN104657713A (en) | Three-dimensional face calibrating method capable of resisting posture and facial expression changes | |
| CN115393548A (en) | Grid texture simplifying algorithm suitable for three-dimensional reconstruction | |
| CN107093182B (en) | A kind of human height's estimation method based on feature corners | |
| CN106931969A (en) | A kind of robot three-dimensional navigation map generation method based on Kinect | |
| CN109727255A (en) | A kind of three-dimensional model building dividing method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170707 |