CN108230247B - Generation method, device, equipment and the computer-readable storage medium of three-dimensional map based on cloud - Google Patents
Generation method, device, equipment and the computer-readable storage medium of three-dimensional map based on cloud Download PDFInfo
- Publication number
- CN108230247B CN108230247B CN201711479953.0A CN201711479953A CN108230247B CN 108230247 B CN108230247 B CN 108230247B CN 201711479953 A CN201711479953 A CN 201711479953A CN 108230247 B CN108230247 B CN 108230247B
- Authority
- CN
- China
- Prior art keywords
- point
- converge
- coordinate system
- converged
- mapping relations
- 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
- G06T3/00—Geometric image transformation in the plane of the image
- G06T3/40—Scaling the whole image or part thereof
- G06T3/4038—Scaling the whole image or part thereof for image mosaicing, i.e. plane images composed of plane sub-images
-
- 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
-
- G—PHYSICS
- G09—EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
- G09B—EDUCATIONAL OR DEMONSTRATION APPLIANCES; APPLIANCES FOR TEACHING, OR COMMUNICATING WITH, THE BLIND, DEAF OR MUTE; MODELS; PLANETARIA; GLOBES; MAPS; DIAGRAMS
- G09B29/00—Maps; Plans; Charts; Diagrams, e.g. route diagram
- G09B29/003—Maps
- G09B29/005—Map projections or methods associated specifically therewith
Abstract
The embodiment of the invention provides a kind of generation method of three-dimensional map based on cloud, device, equipment and computer-readable storage mediums, wherein, the generation method of the three-dimensional map based on cloud includes: and obtains to converge for first point of same building object part exterior wall profile acquisition to converge with second point;Preset mapping relations and corresponding points establish majorized function to data between the coordinate system converged according to first point of coordinate system converged and the second point;The optimal solution of the mapping relations is obtained according to the majorized function;It is converged according to the optimal solution and the second point and converges the outdoor scene three-dimensional map for generating the building corresponding portion with described first point.Using the solution of the present invention, the unification of indoor outdoor scene three-dimensional map and outdoor GPS geographic coordinate system can be completed, realizes and high-precision fusion is carried out to indoor and outdoor three-dimensional map, generates the unified indoor outdoor scene three-dimensional map of indoor and outdoor coordinate system.
Description
Technical field
The present invention relates to three-dimensional map enhancement and treatment fields, can be applied to cloud robot field, can also answer
For robot, unmanned plane avoidance, guide field.In particular to the generation method of a kind of three-dimensional map based on cloud, device,
Equipment and computer-readable storage medium.
Background technique
Positioning outdoor at present with have for GPS (Global Positioning System, global positioning system) compared with
Mature development can reconstruct the outdoor reality that a cloud is dense, texture is fine in conjunction with the location information combination photogrammetric technology of GPS
Scape threedimensional model, and can be unified with GPS geographical co-ordinate system.And when GPS is used for indoor positioning, due to building
Block so that GPS signal be deteriorated, cause indoor positioning accuracy lower, so that indoor outdoor scene three-dimensional map can not be melted
It closes on outdoor outdoor scene three-dimensional map, also cannot achieve the unification of indoor outdoor scene three-dimensional map and GPS geographical co-ordinate system.
In order to realize that indoor outdoor scene three-dimensional map and GPS geographical co-ordinate system are unified, the scheme that the prior art uses is:
When outdoor scene three-dimensional map creates indoors, the fixed road sign of certain amount is set in advance, and obtains the GPS letter of these road signs
Breath, the position where selecting road sign in three-dimensional map model manually are used for the alignment of point cloud.
Although the unification of indoor outdoor scene three-dimensional map and GPS geographical co-ordinate system, but operating cost may be implemented in above scheme
When it is laborious, positioning accuracy is limited.
Summary of the invention
In view of this, an embodiment of the present invention is intended to provide a kind of generation method of three-dimensional map based on cloud, device, setting
Standby and computer-readable storage medium.
First aspect of the embodiment of the present invention provides the method for the generation of the three-dimensional map based on cloud, includes the following steps:
It obtains to converge for first point of same building object part exterior wall profile acquisition and be converged with second point;Described first
Put the three dimensional point cloud collection converged be from the partial building exterior wall profile of taken outdoors;The second point converges
From the three dimensional point cloud collection of the partial building exterior wall profile of indoor acquisition;
Preset mapping relations between the coordinate system converged according to first point of coordinate system converged and the second point
Majorized function is established to data with corresponding points;The corresponding points to be according to described first point converge in point and with its mapping point
The second point that distance is less than first threshold converge in point composition, the mapping point is to converge at described first point
Point is mapped to the second point and converges obtained in coordinate system;
The optimal solution of the mapping relations is obtained according to the majorized function;
It is converged to converge with described first point according to the optimal solution and the second point and generates the building corresponding portion
Outdoor scene three-dimensional map.
Second aspect of the embodiment of the present invention provides a kind of device that three-dimensional map based on cloud generates
Module is obtained, first point for obtaining for same building object part exterior wall profile acquisition converges and second point
It converges;It is three dimensional point cloud collection from the partial building exterior wall profile of taken outdoors that described first point, which is converged,;Institute
Stating second point and converging is three dimensional point cloud collection from the partial building exterior wall profile of indoor acquisition;
Optimization module, it is pre- between the coordinate system for being converged according to first point of coordinate system converged and the second point
If mapping relations and corresponding points majorized function is established to data;The corresponding points to be according to described first point converge in point
And the second point that with its mapping point distance is less than first threshold converge in point composition, the mapping point is by described the
Any converge in point be mapped to the second point and converge obtained in coordinate system;The mapping is obtained according to the majorized function
The optimal solution of relationship;
Generation module is built described in generation for converging to converge with described first point according to the optimal solution and the second point
Build the outdoor scene three-dimensional map of object corresponding portion.
The third aspect of the embodiment of the present invention provides a kind of electronic equipment based on cloud, comprising: communication module, storage
Device, one or more processors;And one or more modules, one or more of modules are stored in the memory
In, and be configured to be executed by one or more of processors, one or more of modules include for executing above-mentioned
On the one hand the instruction of each step in the method provided.
Fourth aspect of the embodiment of the present invention provide with the above-mentioned computer-readable storage medium based on cloud, it is described
Application program includes the computer program being embedded in computer-readable storage medium, and the computer program includes for making
The electronic equipment based on cloud executes the instruction of each step in the method that above-mentioned first aspect provides.
It has the beneficial effect that:
In view of the deficiencies of the prior art, the scheme that the three-dimensional map provided by the embodiment of the present invention based on cloud generates,
It is converged by obtaining to converge for first point of same building object part exterior wall profile acquisition with second point;First cloud
Collection is the three dimensional point cloud collection from the partial building exterior wall profile of taken outdoors;It is from room that the second point, which converges,
The three dimensional point cloud collection of the partial building exterior wall profile of interior acquisition;According to first point of coordinate system converged and
Preset mapping relations and corresponding points establish majorized function to data between the coordinate system that the second point converges;The corresponding points
To be according to described first point converge in point and the second point that is less than first threshold with its mapping point distance converge
Point composition, the mapping point be by described first point converge in point be mapped to the second point and converge in coordinate system and obtain
's;The optimal solution of the mapping relations is obtained according to the majorized function;According to the optimal solution and the second point converge with
Described first point is converged the outdoor scene three-dimensional map for generating the building corresponding portion, can complete indoor outdoor scene three-dimensional map
It with the unification of outdoor GPS geographic coordinate system, realizes and high-precision fusion is carried out to indoor and outdoor three-dimensional map, generate indoor and outdoor coordinate
The outdoor scene three-dimensional map of the indoor and outdoor fusion of system one.
Detailed description of the invention
Fig. 1 is the method flow schematic diagram that the three-dimensional map based on cloud provided in the embodiment of the present invention generates;
Fig. 2 is the apparatus structure schematic diagram that the three-dimensional map based on cloud provided in the embodiment of the present invention generates;
Fig. 3 is the structural schematic diagram of the electronic equipment based on cloud provided in the embodiment of the present invention;
Fig. 4 is the embodiment schematic diagram for the method that the three-dimensional map based on cloud provided in the embodiment of the present invention generates
One;
Fig. 5 is the embodiment schematic diagram for the method that the three-dimensional map based on cloud provided in the embodiment of the present invention generates
Two;
Fig. 6 is the embodiment schematic diagram for the method that the three-dimensional map based on cloud provided in the embodiment of the present invention generates
Three;
Fig. 7 is the embodiment schematic diagram for the method that the three-dimensional map based on cloud provided in the embodiment of the present invention generates
Four.
Specific embodiment
In order to which technical solution of the present invention and advantage is more clearly understood, below in conjunction with attached drawing to of the invention exemplary
Embodiment is described in more detail, it is clear that and the described embodiments are merely a part of the embodiments of the present invention, rather than
The exhaustion of all embodiments.And in the absence of conflict, the feature in the embodiment and embodiment in this explanation can be mutual
It combines.
Fig. 1 is the flow diagram for the method that the three-dimensional map based on cloud provided in the embodiment of the present invention generates, such as
Shown in Fig. 1, the method that should be generated based on the three-dimensional map in cloud be may include steps of:
Step 101 is obtained to converge for first point of same building object part exterior wall profile acquisition and be converged with second point;
It is three dimensional point cloud collection from the partial building exterior wall profile of taken outdoors that described first point, which is converged,;Described second
Put the three dimensional point cloud collection converged be from the partial building exterior wall profile of indoor acquisition;
It is preset between step 102, the coordinate system converged according to first point of coordinate system converged and the second point
Mapping relations and corresponding points establish majorized function to data;The corresponding points to be according to described first point converge in point and with
Its mapping point distance be less than first threshold the second point converge in point composition, the mapping point is by described first point
The point converged is mapped to the second point and converges obtained in coordinate system;
Step 103 obtains the optimal solution of the mapping relations according to the majorized function;
Step 104 converges according to the optimal solution and the second point and converges the generation building with described first point
The outdoor scene three-dimensional map of corresponding portion.
Fig. 4 is the embodiment schematic diagram for the method that the three-dimensional map based on cloud provided in the embodiment of the present invention generates
One, as shown in figure 4, indoor three-dimensional builds the figure stage, by sweeping figure equipment using SLAM (Simultaneous Localization
And Mapping, is positioned at map structuring immediately) etc. build drawing method to the indoor field including the partial building exterior wall
Scape carries out real-time three-dimensional and rebuilds to obtain indoor 3D point cloud, while passing through the GPS number of 3D point cloud in GPS module or equipment collection room
According to 3D point cloud data that figure obtains are built in scanning and its GPS data is correspondingly formed 3D point cloud database, to be subsequently generated interior
It obtains and handles when the unified outdoor scene three-dimensional map of outer positioning.Since the accuracy that GPS is positioned indoors is poor, it can be with
The point cloud data acquired since the outdoor or building body outer surface (includes: that the GPS data of a cloud and SLAM scanning are adopted
The pose data of collection).
It is in an implementation to be easy to those skilled in the art's use/reason since this mode is more commonly used, mainstream using SLAM
Solution, so here by taking SLAM as an example;But in theory, also possible with other modes, it is built as long as can be realized
Scheme the purpose of processing, SLAM is only used for how introduction those skilled in the art specifically implement the present invention, but is not intended to only
SLAM can be used, can be needed in conjunction with practice in implementation process to determine corresponding mode.
Here first point converges the three-dimensional that can be from the outdoor partial building exterior wall profile using GPS gathers
The GPS data collection of point cloud, second point converge the three-dimensional point cloud that can be from the partial building exterior wall profile of indoor acquisition
Visual scanning build figure and obtain data set (also may include the GPS data of three-dimensional point cloud).
Fig. 5 is the embodiment schematic diagram for the method that the three-dimensional map based on cloud provided in the embodiment of the present invention generates
Two, as shown in figure 5, calculating the posture (Rvi, Tvi) for sweeping each position that figure equipment is passed by according to SLAM, and obtain
Corresponding GPS coordinate (xi, yi).First point converge with second point converge in point cloud data can include pose
Data and GPS data, according to first point converge at least two point datas, alternatively, according to second point converge at least two point datas,
Alternatively, according to from first point converge with second point converge at least two point datas (having pose data and GPS data), obtain
First point is converged coordinate system and second point converges the initial mapping relationship (Rm, Tm) of coordinate system, which is rough
's.
Majorized function is established to data further according to the initial mapping relationship and corresponding points.First point converge in point DSLAM,
The mapping point under the coordinate system converged to second point is converted by initial mapping relationship (Rm, Tm) as DSLAMGPS=Rm DSLAM+Tm。
So, in second point converges with mapping point DSLAMGPSPoint of the distance less than first threshold TH converged with above-mentioned first point
Point DSLAMPartner corresponding points pair.Wherein, the specific value of first threshold TH, those skilled in the art can be according to practical feelings
Condition is empirically determined, is not especially limited here.
Determine that a certain number of corresponding points to rear, are established excellent according to corresponding points pair and initial mapping relationship according to the above method
Change function, solves the optimal solution of mapping relations, that is, obtain the mapping relations for making indoor exterior point cloud Accurate align.According to it is described most
Excellent solution and the second point converge the outdoor scene three-dimensional map for converging with described first point and generating the building corresponding portion.
The method that three-dimensional map provided by the embodiment of the present invention based on cloud generates is directed to same building by obtaining
First point of object part exterior wall profile acquisition is converged to converge with second point;It is from described in taken outdoors that described first point, which is converged,
The three dimensional point cloud collection of partial building exterior wall profile;It is from the building portion of indoor acquisition that the second point, which converges,
The exceptionally three dimensional point cloud collection of wall contour;The coordinate converged according to first point of coordinate system converged and the second point
Preset mapping relations and corresponding points establish majorized function to data between system;The corresponding points are to being according to first cloud
The point of concentration and with its mapping point distance be less than first threshold the second point converge in point composition, the mapping point is
By described first point converge in point be mapped to the second point and converge obtained in coordinate system;It is obtained according to the majorized function
The optimal solution of the mapping relations;It is converged to converge with described first point according to the optimal solution and the second point and be built described in generation
The outdoor scene three-dimensional map for building object corresponding portion can complete the GPS geographic coordinate system of indoor outdoor scene three-dimensional map and outdoor
It is unified, it realizes and high-precision fusion is carried out to indoor and outdoor three-dimensional map, generate the reality of the unified indoor and outdoor fusion of indoor and outdoor coordinate system
Scape three-dimensional map.
In implementation, the preset mapping relations can be converges or the second point cloud is concentrated to according to described first point
The data of few two pairs of corresponding points determine by described first point converge in point data map to the second point and converge coordinate system
Under mapping relations, alternatively, according to described first point converge or the second point converge at least two pairs corresponding points data it is true
It is fixed the second point is converged in point data map to first point of mapping relations converged under coordinate system.
As described above, specific implementation in, calculated according to SLAM sweep each position that figure equipment is passed by posture (Rvi,
), and the obtained coordinate (xi, yi) of corresponding GPS Tvi.First point converge with second point converge in point cloud data
It can include pose data and GPS data.At least two pairs of matched points are only needed, first point of coordinate converged can be calculated
The mapping relations (Rm, Tm) between coordinate system that system and the second point converge.It therefore, can be according to first in specific implementation
Point converge at least two point datas, alternatively, according to second point converge at least two point datas, alternatively, being converged according to from first point
With second point converge at least two point datas (having pose data and GPS data), obtain at first point and converge coordinate system and
2 points of initial mapping relationships (Rm, Tm) for converging coordinate system, which is rough.
In implementation, preset between the coordinate system converged according to first point of coordinate system converged and the second point
Mapping relations and corresponding points majorized function is established to data, can specifically include:
It is converged from described first point or the second point converges and obtains several points data;
The several points, which are obtained, according to the preset mapping relations and the several points data is mapped to described
2 points are converged coordinate system or first point of data converged under coordinate system;
It is mapped to it that second point converges coordinate system or converge coordinate according to several point datas at described first point
Data under system, the expression function for establishing Euclidean distance quadratic sum mean value is majorized function.
In specific implementation, first point acquired converge including reference chamber outdoor scene 3D point cloud data and second point cloud
The indoor threedimensional model point cloud data that collection includes, since indoor data are there are error, first point obtained above is converged coordinate
System and second point converge the initial mapping relationship (Rm, Tm) of coordinate system, only converge coordinate system and second point converges seat at first point
One rough mapping relations of mark system therefore can be according to above-mentioned initial mapping relationship (Rm, Tm) and first point of pose converged
Or other data, it maps that under the GPS coordinate system that first point is converged, realizes the system of indoor and outdoor 3D point cloud data (coordinate system)
One.
Such as:
First point converge in point DSLAM, the coordinate system converged to second point is converted by initial mapping relationship (Rm, Tm)
Under mapping point be DSLAMGPS=Rm DSLAM+Tm。
The point cloud D of the coarse data obtained hereinSLAMGPSHave been mapped under GPS coordinate system, obtained initial interior,
The fusion results of outer 3D point cloud data (coordinate system).Based on original fusion obtained above as a result, using ICP (Iterative
Closest Point, iteration closest approach algorithm) method, according to outdoor building outer profile 3D point cloud data DGPSWith it is indoor
Exterior wall profile 3D point cloud data DSLAMGPSAccuracy registration is carried out, one objective function of a minimum is established, specific as follows:
Wherein pt∈DGPS, ps∈DSLAMGPSIt is a pair of of corresponding points pair, a total of NpTo corresponding points.This objective function is practical
The expression function for being above exactly the quadratic sum of the Euclidean distance between all corresponding points is majorized function.
It is described that the optimal solution of the mapping relations is obtained according to the majorized function in implementation, it can specifically include:
Determined according to the majorized function when making the Euclidean distance quadratic sum mean value minimum corresponding mapping relations compared with
Excellent solution;
Converged again from described first point with the second point converge acquisition in addition several pairs of corresponding points data are iterated
Several times to the majorized function, until corresponding mapping relations are described reflect when the variable quantity of the mapping relations is less than threshold value
Penetrate the optimal solution of relationship.
In specific implementation, obtains at first point and converge and converge the other several pairs of corresponding points of acquisition to data point with the second point
ptAnd ps.Assuming that in the case where a mapping relations initial value (R0, T0), to DSLAMGPSIt is converted, one obtained is transformed
Point cloud NDSLAMGPS-0.Then by the point cloud and D after this mapping transformationGPSIt is compared, as long as there are distances in two clouds
Less than certain threshold value TH, it is exactly corresponding points that we, which are considered as the two points,.
There are corresponding points to data ptAnd ps.Rotation and translation (R0, T0) can be estimated with this corresponding points.This
In only 6 freedom degrees in R0 and T0, and the solution of the methods of least square can be used at least more than 6 groups or more for corresponding points quantity
Optimal rotational translation matrix R1 and T1.
Obtain above-mentioned R1 and T1, NDSLAMGPS-0Bring formula into
NDSLAMGPS-1=R1xNDSLAMGPS-0+T1
According to the above-mentioned ND newly obtainedSLAMGPS-1Second point converge in some cloud point mapping transformations after position become
Change, some closest points to also changing accordingly.Therefore, it is necessary to find the corresponding points pair of closest point again.Iteration
It carries out repeatedly, until meeting some stopping criterion for iteration, such as the variable quantity of Rk, Tk are less than certain value or above-mentioned target letter
Several variations is less than certain value or neighbor point to no longer variation etc..
By above-mentioned ICP algorithm can by include indoor threedimensional model point cloud data second point converge coordinate system and
First point including outdoor outdoor scene threedimensional model point cloud data is converged coordinate system and is merged, and global unified indoor and outdoor three is obtained
Tie up map.
In relocation phase, it directly can be obtained by outdoor position and orientation using GPS and magnetometer outdoor, at this time
The outdoor threedimensional model of 3D high-precision is loaded, and obtains matched visual angle picture.It, can be using the view such as SLAM if going to interior
Feel locating scheme, the pose data that SLAM is handled is mapped to outdoor GPS coordinate data according to mapping relations, and count
The three-dimensional rendering figure under current visual angle is calculated, does not have to switch 3D threedimensional model because GPS signal is impacted to realize, carries out nothing
The 3D of seam is browsed.
Based on the same inventive concept, a kind of dress that the three-dimensional map based on cloud generates is additionally provided in the embodiment of the present invention
It sets, since the principle that these equipment solve the problems, such as is similar to the method that the above-mentioned three-dimensional map based on cloud generates, these
The implementation of equipment may refer to the implementation of method, and overlaps will not be repeated.
Fig. 2 is the structural schematic diagram for the device that the three-dimensional map based on cloud provided in the embodiment of the present invention generates, such as
Shown in Fig. 2, the device that should be generated based on the three-dimensional map in cloud may include:
Module 201 is obtained, first point for obtaining for same building object part exterior wall profile acquisition converges and the
2 points are converged;It is three dimensional point cloud from the partial building exterior wall profile of taken outdoors that described first point, which is converged,
Collection;It is three dimensional point cloud collection from the partial building exterior wall profile of indoor acquisition that the second point, which converges,;
Optimization module 202, coordinate system for being converged according to first point of coordinate system converged and the second point it
Between preset mapping relations and corresponding points majorized function is established to data;The corresponding points are to being converged according to described first point
Point and with its mapping point distance be less than first threshold the second point converge in point composition, the mapping point is by institute
State first point converge in point be mapped to the second point and converge obtained in coordinate system;It is obtained according to the majorized function described
The optimal solution of mapping relations;
Generation module 203 generates institute for converging to converge with described first point according to the optimal solution and the second point
State the outdoor scene three-dimensional map of building corresponding portion.
In implementation, during the optimization module 202 is further used for being converged according to described first point or the second point converges
The data of at least two pairs corresponding points determine by described first point converge in point data map to the second point and converge coordinate
System under mapping relations, alternatively, according to described first point converge or the second point converge at least two pairs corresponding points data
It is determining using the second point converge in point data map to first point of mapping relations converged under coordinate system as institute
State preset mapping relations.
In implementation, if the optimization module 202 is specifically used for converging from described first point or the second point converges acquisition
Dry point data;According to the preset mapping relations and the several points data obtain the several points be mapped to it is described
Second point converges coordinate system or first point of data converged under coordinate system;Institute is mapped to it according to several point datas
It states second point and converges coordinate system or first point of data converged under coordinate system, establish the expression of Euclidean distance quadratic sum mean value
Function is majorized function.
In implementation, the optimization module 202 is specifically used for making the Euclidean distance square according to majorized function determination
Corresponding mapping relations are more excellent solution when with mean value minimum;It is converged again from described first point and converges acquisition in addition with the second point
Several pairs of corresponding points are iterated to the majorized function several times data, until the variable quantity of the mapping relations is less than threshold
Corresponding mapping relations are the optimal solution of the mapping relations when value.
Based on the same inventive concept, a kind of electronic equipment based on cloud is additionally provided in the embodiment of the present invention, Fig. 3 is this
The structural schematic diagram of the electronic equipment based on cloud provided in inventive embodiments, as shown in figure 3, electronics of this kind based on cloud
Equipment 300 may include: communication module 301, memory 302, one or more processors 303;And one or more modules,
One or more of modules are stored in the memory, and are configured to be executed by one or more of processors,
One or more of modules include the finger for executing each step in the method that the above-mentioned three-dimensional map based on cloud generates
It enables.
Based on the same inventive concept, additionally provided in the embodiment of the present invention it is a kind of with above-mentioned based on the computer-readable of cloud
Storage medium, the computer-readable storage medium includes the computer journey being embedded in computer-readable storage medium
Sequence, the computer program include raw for making the electronic equipment based on cloud execute the above-mentioned three-dimensional map based on cloud
At method in each step instruction.
Embodiment 1
Fig. 6 is the embodiment schematic diagram for the method that the three-dimensional map based on cloud provided in the embodiment of the present invention generates
Three, as shown in fig. 6, three-dimensional acquisition equipment executes two pieces while obtaining GPS data and appoints when carrying out building figure to indoor scene
Business: first is that with the picture of RGB camera acquisition current visual angle, it is transmitted to vSLAM (visual Simultaneous
Localization And Mapping, view-based access control model are positioned at map structuring immediately) algorithm is to carrying out interior space feature
Spot scan is simultaneously positioned to pose data (Rv, Tv) in real time, generates SLAM map;Second is that according to calculated pose data (Rv,
Tv) pass through with the real-time deep information Depth of depth camera acquisition and the three-dimensional map on operations progress backstage such as merge, denoise and build
Mould generates indoor three-dimensional point and converges data, i.e., above-mentioned first point is converged data.
According to the GPS data for having recorded three-dimensional acquisition equipment, the indoor coordinate system and outdoor seat of this scene rebuilding are calculated
Mark the mapping relations of system, that is, converge coordinate system and second point converges the initial mapping relationship of coordinate system at above-mentioned first point.Assuming that choosing
The GPS data of a road sign point of i (i >=2) is taken (or to arrange the road sign of four accurate positionings with the mode surveyed and drawn indoors in advance
Point), obtain i GPS data (x1, y1)~(xi, yi) and corresponding indoor map vSLAM pose data (Rv1, Tv1)~
(Rvi, Tvi), calculate indoor coordinate system to outdoor coordinate system mapping relations (Rm, Tm).
Indoor high-precision three-dimensional point cloud can be roughly mapped in outdoor outdoor scene three-dimensional global map and be obtained according to (Rm, Tm)
To initial fusion results.It is then based on above-mentioned initial fusion results and obtains mapping relations most using accurate ICP algorithm
Excellent solution (Rn, Tn), then according to the optimal solution of mapping relations (Rn, Tn) by indoor threedimensional model and existing outdoor outdoor scene three
Dimension module carries out high-precision, seamless fusion.Detailed description has been carried out in detailed process above, and which is not described herein again.
When carrying out outdoor positioning, two-dimensional localization information (x, y) is directly obtained using GPS.When from the outside into the room,
RGB-vSLAM positioning is switched to from GPS positioning.The vSLAM space characteristics point location library for building that the figure stage builds up is used, forward sight will be worked as
The image of angle acquisition is matched with the space characteristics point in library, calculates and optimize the spin matrix Rv and displacement square of current visual angle
The three-dimensional localization of indoor part can be obtained in battle array Tv.Calculate before using again indoor positioning to initial mapping relations (Rm,
Tm) and (Rn, Tn), indoor three-dimensional pose data (Rv, Tv) can be converted to outdoor two-dimensional localization (x, y), thus real
Existing indoor and outdoor positioning is unified.Meanwhile the three-dimensional point off density cloud built up in the preparation stage can also be used when indoor navigation, indoors
Real-time avoidance navigation is realized when navigation.
Embodiment 2
Fig. 7 is the embodiment schematic diagram for the method that the three-dimensional map based on cloud provided in the embodiment of the present invention generates
Four, as shown in fig. 7, indoor three-dimensional builds figure and locating scheme can be for using laser radar and RGB camera, difference be only that room
Interior three-dimensional builds figure and positioning using device type different from, and the method and step being embodied is identical with above-described embodiment 1,
Which is not described herein again.
The embodiment of the present invention proposes a kind of indoor and outdoor three-dimensional high-precision map automatic alignment method, and is regarded according to SLAM etc.
Feel the reorientation for building nomography and mapping relations for high-precision 3D point cloud figure real-time rendering to current visual angle, realize high-precision,
Outdoor scene is digitized clear
It should be understood by those skilled in the art that, the embodiment of the present invention can provide as method, system or application program.
Therefore, complete hardware embodiment, complete software embodiment or embodiment combining software and hardware aspects can be used in the present invention
Form.It is deposited moreover, the present invention can be used to can be used in the computer that one or more wherein includes computer usable program code
The form for the application program implemented on storage media (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.).
The present invention be referring to according to the method for the embodiment of the present invention, the flow chart of equipment (system) and application program and/or
Block diagram describes.It should be understood that each process that can be realized by computer program instructions in flowchart and/or the block diagram and/or
The combination of process and/or box in box and flowchart and/or the block diagram.It can provide these computer program instructions to arrive
General purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices processor to generate one
Machine, so that being generated by the instruction that computer or the processor of other programmable data processing devices execute for realizing flowing
The device for the function of being specified in journey figure one process or multiple processes and/or block diagrams one box or multiple boxes.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
Although preferred embodiments of the present invention have been described, it is created once a person skilled in the art knows basic
Property concept, then additional changes and modifications may be made to these embodiments.So it includes excellent that the following claims are intended to be interpreted as
It selects embodiment and falls into all change and modification of the scope of the invention.
Claims (10)
1. a kind of method that the three-dimensional map based on cloud generates, which comprises the steps of:
It obtains to converge for first point of same building object part exterior wall profile acquisition and be converged with second point;First cloud
Collection is the three dimensional point cloud collection from the partial building exterior wall profile of taken outdoors;It is from room that the second point, which converges,
The three dimensional point cloud collection of the partial building exterior wall profile of interior acquisition;
Between the coordinate system converged according to first point of coordinate system converged and the second point preset mapping relations with it is right
It should put and majorized function is established to data;The corresponding points to be according to described first point converge in point and with its mapping point distance
Less than first threshold the second point converge in point composition, the mapping point be by described first point converge in point reflect
The second point is mapped to converge obtained in coordinate system;
The optimal solution of the mapping relations is obtained according to the majorized function;
It is converged according to the optimal solution and the second point and converges the reality for generating the building corresponding portion with described first point
Scape three-dimensional map.
2. the method that the three-dimensional map based on cloud generates as described in claim 1, which is characterized in that the preset mapping
Relationship be according to described first point converge or the second point converge at least two pairs corresponding points data determine by described the
Any converge in point data map to the second point and converge mapping relations under coordinate system, alternatively, according to described first point
Converge or the second point converge at least two pairs corresponding points data determine the second point is converged in point data reflect
It is incident upon first point of mapping relations converged under coordinate system.
3. the method that three-dimensional map based on cloud generates as described in claim 1, which is characterized in that described according to described the
Preset mapping relations and corresponding points establish data between the coordinate system that some coordinate system converged and the second point converge
Majorized function specifically includes:
It is converged from described first point or the second point converges and obtains several points data;
The several points, which are obtained, according to the preset mapping relations and the several points data is mapped to the second point
Converge coordinate system or first point of data converged under coordinate system;
It is mapped to it that second point converges coordinate system or converge under coordinate system according to several point datas at described first point
Data, establish Euclidean distance quadratic sum mean value expression function be majorized function.
4. the method that the three-dimensional map based on cloud generates as claimed in claim 3, which is characterized in that described according to described excellent
Change function and obtain the optimal solution of the mapping relations, specifically includes:
Corresponding mapping relations more excellent solution when making the Euclidean distance quadratic sum mean value minimum is determined according to the majorized function;
Again from described first point converge with the second point converge acquisition in addition several pairs of corresponding points data are iterated to institute
Majorized function is stated several times, until it is mapping pass that the variable quantity of the mapping relations, which is less than corresponding mapping relations when threshold value,
The optimal solution of system.
5. a kind of device that the three-dimensional map based on cloud generates characterized by comprising
Module is obtained, first point for obtaining for same building object part exterior wall profile acquisition converges and second point cloud
Collection;It is three dimensional point cloud collection from the partial building exterior wall profile of taken outdoors that described first point, which is converged,;It is described
It is three dimensional point cloud collection from the partial building exterior wall profile of indoor acquisition that second point, which converges,;
Optimization module, it is preset between the coordinate system for being converged according to first point of coordinate system converged and the second point
Mapping relations and corresponding points establish majorized function to data;The corresponding points to be according to described first point converge in point and with
Its mapping point distance be less than first threshold the second point converge in point composition, the mapping point is by described first point
The point converged is mapped to the second point and converges obtained in coordinate system;The mapping relations are obtained according to the majorized function
Optimal solution;
Generation module generates the building for converging to converge with described first point according to the optimal solution and the second point
The outdoor scene three-dimensional map of corresponding portion.
6. the device that three-dimensional map based on cloud generates as claimed in claim 5, which is characterized in that the optimization module into
One step is used to be converged according to described first point or the second point converge at least two pairs corresponding points data determine will be described
First point converge in point data map to the second point and converge mapping relations under coordinate system, alternatively, according to described first
Point converges or the second point converge at least two pairs corresponding points data determine the second point is converged in point data
First point of mapping relations converged under coordinate system are mapped to as the preset mapping relations.
7. the device that the three-dimensional map based on cloud generates as claimed in claim 5, which is characterized in that the optimization module tool
Body is used to converge from described first point or the second point converges and obtains several points data;According to the preset mapping relations
And the several points data obtain the several points and are mapped to that the second point converges coordinate system or converge at described first point
Data under coordinate system;The second point, which is mapped to, with it according to several point datas converges coordinate system or first cloud
Collect the data under coordinate system, the expression function for establishing Euclidean distance quadratic sum mean value is majorized function.
8. the device that the three-dimensional map based on cloud generates as claimed in claim 7, which is characterized in that the optimization module tool
Body is used to determine that corresponding mapping relations are more excellent when making the Euclidean distance quadratic sum mean value minimum according to the majorized function
Solution;Again from described first point converge with the second point converge acquisition in addition several pairs of corresponding points data are iterated to described
Several times, corresponding mapping relations are the mapping relations to majorized function when the variable quantity of the mapping relations is less than threshold value
Optimal solution.
9. a kind of electronic equipment based on cloud characterized by comprising communication module, memory, one or more processing
Device;And one or more modules, one or more of modules are stored in the memory, and are configured to by described
One or more processors execute, and one or more of modules include requiring any the method in 1 to 4 for perform claim
In each step instruction.
10. a kind of computer-readable storage medium, the computer-readable storage medium is computer-readable including being embedded in
Storage medium in computer program, the computer program include for perform claim require 1 to 4 in any the method
In each step instruction.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711479953.0A CN108230247B (en) | 2017-12-29 | 2017-12-29 | Generation method, device, equipment and the computer-readable storage medium of three-dimensional map based on cloud |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711479953.0A CN108230247B (en) | 2017-12-29 | 2017-12-29 | Generation method, device, equipment and the computer-readable storage medium of three-dimensional map based on cloud |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108230247A CN108230247A (en) | 2018-06-29 |
CN108230247B true CN108230247B (en) | 2019-03-15 |
Family
ID=62646203
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711479953.0A Active CN108230247B (en) | 2017-12-29 | 2017-12-29 | Generation method, device, equipment and the computer-readable storage medium of three-dimensional map based on cloud |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108230247B (en) |
Families Citing this family (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109636917B (en) * | 2018-11-02 | 2023-07-18 | 北京微播视界科技有限公司 | Three-dimensional model generation method, device and hardware device |
CN109543634B (en) * | 2018-11-29 | 2021-04-16 | 达闼科技(北京)有限公司 | Data processing method and device in positioning process, electronic equipment and storage medium |
CN109903382B (en) * | 2019-03-20 | 2023-05-23 | 中煤航测遥感集团有限公司 | Point cloud data fusion method and device |
CN110288650B (en) * | 2019-05-27 | 2023-02-10 | 上海盎维信息技术有限公司 | Data processing method and scanning terminal for VSLAM |
CN111080518B (en) * | 2019-11-27 | 2023-10-20 | 贝壳技术有限公司 | Point cloud image stitching method and device and storage medium |
CN111080682B (en) * | 2019-12-05 | 2023-09-01 | 北京京东乾石科技有限公司 | Registration method and device for point cloud data |
CN111473756B (en) * | 2020-04-03 | 2021-07-06 | 上海嘉奥信息科技发展有限公司 | Surface curve measuring method, system and medium based on Unity3D model |
CN112669250B (en) * | 2021-03-16 | 2021-09-17 | 湖北亿咖通科技有限公司 | Track alignment method and electronic equipment |
CN113066112B (en) * | 2021-03-25 | 2021-10-22 | 泰瑞数创科技(北京)有限公司 | Indoor and outdoor fusion method and device based on three-dimensional model data |
CN113409459B (en) * | 2021-06-08 | 2022-06-24 | 北京百度网讯科技有限公司 | Method, device and equipment for producing high-precision map and computer storage medium |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104807460A (en) * | 2015-05-04 | 2015-07-29 | 深圳大学 | Indoor positioning method and system for unmanned aerial vehicle |
CN105702151A (en) * | 2016-03-31 | 2016-06-22 | 百度在线网络技术(北京)有限公司 | Indoor map constructing method and device |
CN105828296A (en) * | 2016-05-25 | 2016-08-03 | 武汉域讯科技有限公司 | Indoor positioning method based on convergence of image matching and WI-FI |
CN106646562A (en) * | 2016-09-09 | 2017-05-10 | 华东师范大学 | High-precision three-dimensional real scene indoor and outdoor integrated positioning method and device |
CN106780699A (en) * | 2017-01-09 | 2017-05-31 | 东南大学 | A kind of vision SLAM methods aided in based on SINS/GPS and odometer |
CN106971403A (en) * | 2017-04-27 | 2017-07-21 | 武汉数文科技有限公司 | Point cloud chart is as processing method and processing device |
Family Cites Families (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9760996B2 (en) * | 2015-08-11 | 2017-09-12 | Nokia Technologies Oy | Non-rigid registration for large-scale space-time 3D point cloud alignment |
US10066925B2 (en) * | 2016-02-02 | 2018-09-04 | The Boeing Company | Point cloud processing apparatus and method |
CN106639323B (en) * | 2016-10-12 | 2023-01-24 | 上海建工集团股份有限公司 | Building reconstruction system and method based on three-dimensional laser scanning and 3D printing |
CN106530345B (en) * | 2016-11-07 | 2018-12-25 | 江西理工大学 | A kind of building three-dimensional laser point cloud feature extracting method under same machine Image-aided |
-
2017
- 2017-12-29 CN CN201711479953.0A patent/CN108230247B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104807460A (en) * | 2015-05-04 | 2015-07-29 | 深圳大学 | Indoor positioning method and system for unmanned aerial vehicle |
CN105702151A (en) * | 2016-03-31 | 2016-06-22 | 百度在线网络技术(北京)有限公司 | Indoor map constructing method and device |
CN105828296A (en) * | 2016-05-25 | 2016-08-03 | 武汉域讯科技有限公司 | Indoor positioning method based on convergence of image matching and WI-FI |
CN106646562A (en) * | 2016-09-09 | 2017-05-10 | 华东师范大学 | High-precision three-dimensional real scene indoor and outdoor integrated positioning method and device |
CN106780699A (en) * | 2017-01-09 | 2017-05-31 | 东南大学 | A kind of vision SLAM methods aided in based on SINS/GPS and odometer |
CN106971403A (en) * | 2017-04-27 | 2017-07-21 | 武汉数文科技有限公司 | Point cloud chart is as processing method and processing device |
Non-Patent Citations (1)
Title |
---|
用于弱纹理场景三维重建的机器人视觉系统;林义闽;《光学 精密工程》;20150228;全文 |
Also Published As
Publication number | Publication date |
---|---|
CN108230247A (en) | 2018-06-29 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108230247B (en) | Generation method, device, equipment and the computer-readable storage medium of three-dimensional map based on cloud | |
Nüchter et al. | 6D SLAM—3D mapping outdoor environments | |
Moghadam et al. | Line-based extrinsic calibration of range and image sensors | |
CN106767827B (en) | Mobile robot point cloud map creation method based on laser data | |
CN112927360A (en) | Three-dimensional modeling method and system based on fusion of tilt model and laser point cloud data | |
Jessup et al. | Robust and efficient multirobot 3-d mapping merging with octree-based occupancy grids | |
Borrmann et al. | The project thermalmapper–thermal 3d mapping of indoor environments for saving energy | |
US20040239756A1 (en) | Method and apparatus for computing error-bounded position and orientation of panoramic cameras in real-world environments | |
CN104463969B (en) | A kind of method for building up of the model of geographical photo to aviation tilt | |
Li et al. | Leveraging planar regularities for point line visual-inertial odometry | |
Fossel et al. | 2D-SDF-SLAM: A signed distance function based SLAM frontend for laser scanners | |
CN109754459A (en) | It is a kind of for constructing the method and system of human 3d model | |
Jessup et al. | Robust and efficient multi-robot 3D mapping with octree based occupancy grids | |
Pitzer et al. | Automatic reconstruction of textured 3D models | |
Koch et al. | Wide-area egomotion estimation from known 3d structure | |
Dold | Extended Gaussian images for the registration of terrestrial scan data | |
Ishikawa et al. | In-situ 3d indoor modeler with a camera and self-contained sensors | |
Ma et al. | Location and 3-D visual awareness-based dynamic texture updating for indoor 3-D model | |
CN112132876B (en) | Initial pose estimation method in 2D-3D image registration | |
Ruchay et al. | Accurate reconstruction of the 3D indoor environment map with a RGB-D camera based on multiple ICP | |
Klein et al. | The view-cube: An efficient method of view planning for 3d modelling from range data | |
Ge et al. | Depth estimation using a sliding camera | |
Feng et al. | Research on Calibration Method of Multi-camera System without Overlapping Fields of View Based on SLAM | |
Julia et al. | Autonomous surveying of underfloor voids | |
Ruan | The survey of vision-based 3D modeling techniques |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |