CN110163968A - RGBD camera large-scale three dimensional scenario building method and system - Google Patents
RGBD camera large-scale three dimensional scenario building method and system Download PDFInfo
- Publication number
- CN110163968A CN110163968A CN201910452208.XA CN201910452208A CN110163968A CN 110163968 A CN110163968 A CN 110163968A CN 201910452208 A CN201910452208 A CN 201910452208A CN 110163968 A CN110163968 A CN 110163968A
- Authority
- CN
- China
- Prior art keywords
- subgraph
- point cloud
- dimensional
- cloud
- algorithm
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
- G06T17/05—Geographic models
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Geometry (AREA)
- Software Systems (AREA)
- Remote Sensing (AREA)
- Computer Graphics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The invention discloses a kind of RGBD camera large-scale three dimensional scenario building method and system, method includes the following steps: obtain the depth image and color image of RGBD camera, three-dimensional point cloud is generated, and by point cloud registering to laser radar referential；The number of scan points evidence for obtaining robot pose data and laser radar, is handled it using the Local Optimization Algorithm of Cartographerd algorithm, spanning subgraph and robot odometer information；According to robot odometer information, the subgraph of generation is bound with corresponding cloud, splicing subgraph generates three-dimensional map.
Description
Technical field
This disclosure relates to dimensional Modeling Technology field, and in particular to a kind of RGBD camera based on Cartographer algorithm
Large-scale three dimensional scenario building method and system.
Background technique
In the intelligent Application of mobile robot, the intelligence degree of robot depends on robot on most and exists
Positioning accuracy in scene, it is therefore desirable to provide the environmental map of pinpoint accuracy for robot.Environmental map generally has two-dimensionally
Figure and two kinds of forms of three-dimensional map are being answered wherein only having plane two-dimensional data using the two-dimensional map that two-dimensional laser radar constructs
With that can lose compared with multi information in the process, it is unfavorable for robot and is navigated and positioned, so three-dimensional map has at many aspects
Extremely important application and research value.
Comparison can generate the three-dimensional laser radar, depth camera and binocular camera of three-dimensional data.Wherein RGBD camera by
Measuring principle, measurement accuracy and the limitation for measuring distance, effect is poor in the application process that large scene builds figure, larger when encountering
Interference or robot operation it is very fast when, be easily lost self-position, although so can pass through change image between constraint close
System solves the problems, such as that position is lost, but all inevasible increases computer resource usage.And compared with RGBD camera, although
The anti-interference ability of three-dimensional laser radar and binocular camera is stronger, but the price is very expensive for three-dimensional laser radar, the point of generation
Cloud is sparse, lacks colouring information；Binocular camera is then to need to expend additional computing resource contrast images calculating depth data, two
Person carry out building indoors figure rapidity and accuracy it is insufficient.Therefore this paper presents a kind of fusion laser radars and RGBD phase
Two kinds of sensors of machine carry out dense build using the pose superposition RGBD camera point cloud of laser SLAM output robot under large scene
The method of figure.
Several laser SLAM algorithms have currently popular: (1) Gmapping algorithm is a kind of based on particle filter after improving
Method builds nomography, and number of particles is more under large scene can expend more resource；(2) Karto algorithm, using figure optimization
Method is adjusted figure instead of particle filter, and using SPA, but can have when large scene builds figure a large amount of
The insertion of landmark can also expend a large amount of memories；(3) Hector algorithm carries out scanning point set using Gauss-Newton method
Scan matching, it is higher to sensor requirements although memory consumption is less, need the laser radar of degree of precision can just build out compared with
Good map；(4) method that Cartographer algorithm uses pose refinement, by the data frame of laser radar and Submap into
Row matching, and global optimization is carried out by finding closed loop constraint in global optimization procedure, compare first three methods, program energy
Enough eliminate most accumulated error, and the laser radar extra high without precision.
It is mostly using VSLAM (visual simultaneous in RGBD camera three-dimensional rebuilding method currently popular
Localization and mapping) algorithm front end calculate RGBD image characteristic point between the constraint relationship and press this relationship
Summing point cloud, while three-dimensional map is constructed in the method that rear end carries out winding optimization to front end data.Inventor is in R&D process
Middle discovery, this method are limited by RGBD camera measurement principle, measurement accuracy and measurement distance, in large scene, environment it is complicated,
Figure precision is poor, is easily lost robot location there are building in the case where a large amount of sunlight interferences, and can expend when calculating a large amount of
Resource needs the computer of superior performance just to can guarantee calculating speed.
Summary of the invention
In order to overcome the above-mentioned deficiencies of the prior art, the disclosure is to be equipped with the Indoor Robot of laser radar Yu RGBD camera
For research object, proposes a kind of RGBD camera large-scale three dimensional scenario building method based on Cartographer algorithm and be
The characteristics of system, strong antijamming capability few by laser radar data, pass through Cartographer laser SLAM algorithm computing machine
People's pose, insertion puts cloud in real time and builds figure in real time to realize in the subgraph that Cartographer is generated；It uses simultaneously
The winding optimization function of Cartographer is adjusted the pose being inserted into before, and experiment shows to compare under reality scene
Simple using laser radar or RGBD camera build figure precision higher, builds figure speed faster.
A kind of RGBD camera large-scale three dimensional scenario building method based on Cartographer algorithm that the disclosure provides
Technical solution is:
A kind of RGBD camera large-scale three dimensional scenario building method based on Cartographer algorithm, this method includes following
Step:
The depth image and color image of RGBD camera are obtained, generates three-dimensional point cloud, and by point cloud registering to laser radar
Referential；
The number of scan points evidence for obtaining robot pose data and laser radar, utilizes the part of Cartographerd algorithm
Optimization algorithm handles it, spanning subgraph and robot odometer information；
According to robot odometer information, the subgraph of generation is bound with corresponding cloud, splicing subgraph generates three
Tie up map.
Further, described cloud generation method are as follows:
RGBD camera is registrated, depth camera referential is registrated under color camera referential；
According to any point in the corresponding calculated for pixel values color image in depth image after registration in space three
Tie up location information；
Whole picture depth image is successively traversed, corresponding cloud of camera is obtained.
Further, described to include: by point cloud registering to the step of laser radar referential
Robot chassis is obtained to enamel deviation of the camera photocentre at a distance from laser radar center and its in yaw angle
Angle；
Converted using coordinate of the transformation matrix to each point in camera point cloud, by the scanning point set of laser radar with
Point cloud is registrated.
Further, described the step of being bound subgraph with corresponding cloud, includes:
The posture information and timestamp of the subgraph generated are recorded, the timestamp generated when by spanning subgraph is adopted with RGBD camera
The timestamp of collection depth image compares, and selects the RGBD camera depth image information on the time near subgraph；
Time synchronization will be carried out near the RGBD camera depth image information of subgraph and robot posture information on time；
The differential seat angle of horizontal displacement difference and rotation when calculating spanning subgraph between robot and subgraph, forms spin moment
Battle array, will be in the referential of Cloud transform to its correspondence subgraph using spin matrix；
The point cloud transformed in subgraph referential and picture information are bound, and a cloud is displaced, it will be transformed
Point cloud moves to Sub-graph Space position.
It further, further include that point cloud bound in the subgraph changed is occurred into pose when subgraph pose changes
The step of being translated again according to subgraph variation.
A kind of RGBD camera large-scale three dimensional scene based on Cartographer algorithm that another aspect of the present disclosure provides
The technical solution of construction method is:
A kind of RGBD camera large-scale three dimensional scenario building system based on Cartographer algorithm, the system include:
Point cloud registering module generates three-dimensional point cloud for obtaining the depth image and color image of RGBD camera, and by point
Cloud is registrated to laser radar referential；
Subgraph generation module is utilized for obtaining the number of scan points evidence of robot pose data and laser radar
The Local Optimization Algorithm of Cartographerd algorithm handles it, spanning subgraph and robot odometer information；
Point cloud binding module, for according to robot odometer information, the subgraph of generation to be tied up with corresponding cloud
It is fixed；
Three-dimensional building module generates three-dimensional map for splicing subgraph；
Winding point cloud module, for when subgraph pose changes, point bound in the subgraph changed to be occurred in pose
Cloud is translated again according to subgraph variation.
The technical solution for the computer readable storage medium that another aspect of the present disclosure provides is:
A kind of computer readable storage medium, is stored thereon with computer program, realization when which is executed by processor
Step in RGBD camera large-scale three dimensional scenario building method based on Cartographer algorithm as described above.
A kind of technical solution for computer equipment that another aspect of the present disclosure provides is:
A kind of computer equipment can be run on a memory and on a processor including memory, processor and storage
Computer program, the processor are realized when executing described program as described above based on the RGBD phase of Cartographer algorithm
Step in machine large-scale three dimensional scenario building method.
Through the above technical solutions, the beneficial effect of the disclosure is:
(1) disclosure is greatly promoted using the big field angle of laser radar and the high robust of Cartographer algorithm
The anti-interference ability and the algorithm speed of service of three-dimensional reconstruction under RGBD camera large scene；
(2) disclosure is using the winding optimization function of Cartographer algorithm to corresponding to the changed subgraph of pose
Point cloud translated again, eliminate the accumulated error built and generated during figure, improve RGBD camera three-dimensional reconstruction map
Precision.
Detailed description of the invention
The Figure of description for constituting a part of this disclosure is used to provide further understanding of the disclosure, and the disclosure is shown
Meaning property embodiment and its explanation do not constitute the improper restriction to the disclosure for explaining the application.
Fig. 1 is the process of RGBD camera large-scale three dimensional scenario building method of the embodiment one based on Cartographer algorithm
Figure one；
Fig. 2 is the process of RGBD camera large-scale three dimensional scenario building method of the embodiment one based on Cartographer algorithm
Figure two；
Fig. 3 (a) is that KinectV2 depth image with RGB image is registrated figure before embodiment one is corrected；
Fig. 3 (b) is that KinectV2 depth image with RGB image is registrated figure after embodiment one is corrected；
Fig. 4 is one Kinect of embodiment and radar point cloud registering figure；
Fig. 5 is the binding schematic diagram of one Submap of embodiment and point cloud data；
Fig. 6 (a) is Cartographer winding effect picture before one winding of embodiment；
Fig. 6 (b) is Cartographer winding effect picture after one winding of embodiment；
Fig. 7 (a) and Fig. 7 (b) is one RGBD camera large-scale three dimensional scenario building method acquired results figure of embodiment；
Fig. 7 (c) and Fig. 7 (d) is RTAB algorithm acquired results figure；
Fig. 7 (e) and Fig. 7 (f) is color image and depth image under sunlight interference；
Fig. 8 (a) and Fig. 8 (b) is one RGBD camera large-scale three dimensional scenario building method acquired results figure of embodiment；
Fig. 8 (c) and Fig. 8 (d) is that the point cloud superposition precision of one RGBD camera large-scale three dimensional scenario building method of embodiment is shown
It is intended to；
Fig. 8 (e) and Fig. 8 (f) is RTAB algorithm acquired results figure；
Fig. 9 is that Cartographer builds figure precision schematic diagram；
Figure 10 (a) and Figure 10 (b) is second floor experimental precision contrast schematic diagram；
Figure 10 (c) and Figure 10 (d) is Stall experimental precision contrast schematic diagram.
Specific embodiment
The disclosure is described further with embodiment with reference to the accompanying drawing.
It is noted that following detailed description is all illustrative, it is intended to provide further instruction to the disclosure.Unless another
It indicates, all technical and scientific terms that the disclosure uses have logical with disclosure person of an ordinary skill in the technical field
The identical meanings understood.
It should be noted that term used herein above is merely to describe specific embodiment, and be not intended to restricted root
According to the illustrative embodiments of the application.As used herein, unless the context clearly indicates otherwise, otherwise singular
Also it is intended to include plural form, additionally, it should be understood that, when in the present specification using term "comprising" and/or " packet
Include " when, indicate existing characteristics, step, operation, device, component and/or their combination.
Embodiment one
The present embodiment proposes one kind and is based on to be equipped with the artificial research object of indoor machine of laser radar Yu RGBD camera
The RGBD camera large-scale three dimensional scenario building method of Cartographer algorithm, anti-interference ability few by laser radar data
Strong feature, by Cartographer laser SLAM algorithm calculating robot's pose, in the subgraph that Cartographer is generated
Middle insertion puts cloud in real time and builds figure in real time to realize.Simultaneously using the winding optimization function of Cartographer to the position being inserted into before
Appearance is adjusted, and experiment shows to carry out building figure precision using laser radar or RGBD camera compared to simple under reality scene
It is higher, build figure speed faster.
Please refer to attached drawing 1, the RGBD camera large-scale three dimensional scenario building method packet based on Cartographer algorithm
Include following steps:
S101 obtains RGBD camera depth image and color image, generates three-dimensional point cloud, and by point cloud registering to laser thunder
Up to referential.
Specifically, the implementation of the step 101 is as follows:
S1011 obtains the depth image and color image of RGBD camera.
Specifically, the depth image data and color image data of RGBD camera Kinectv2 acquisition are obtained.
In order to which the data of RGBD camera Kinectv2 acquisition and the data fusion of laser radar acquisition use, need to pass two
Under sensor Registration of Measuring Data to the same coordinate system.
S1012, point Yun Shengcheng.
In order to obtain accurate point cloud, RGBD camera is registrated first, depth camera referential is registrated to RGB phase
Under machine referential, as shown in Fig. 3 (a) and Fig. 3 (b).
After completing registration, any point (x in RGB imageimg, yimg), it can be according to the phase in depth image after registration
The calculated for pixel values answered goes out the depth value of the point, and can calculate the point in space according to formula (1), (2), (3)
Position p=[px,py,pz]T, traversing whole picture depth image can be obtained corresponding cloud of camera.
Wherein, Z0Indicate the distance from camera to reference planes used when calibration, f indicates that depth camera focal length, b indicate
For the distance of depth camera to RGB camera, (cx, cy) indicate depth camera optical center coordinate.
S1013, the registration of point cloud to laser radar coordinate system.
After obtaining camera point cloud, in order to which the grating map that cloud is accurately inserted into Cartographer generation will be put
In, it is also necessary to cloud is registrated with laser radar scanning point.
Measurement can obtain deviation of the RGB camera optical center at a distance from laser radar center and its on Yaw axis on robot chassis
Angle ξ=(ξx,ξy,ξz,ξθ), because camera and radar relative level are placed, it is possible to ignore Pitch axis and Roll axis
Deviator.After measuring above data, transformation matrix T is usedξThe coordinate of each point in camera point cloud is converted:
Wherein, p is position a little in space.
The scanning point set that laser radar can be completed is registrated with point cloud, and effect picture is as shown in Figure 4.
S102 obtains robot pose and laser radar data, utilizes the Local Optimization Algorithm of Cartographerd algorithm
Obtained data are handled, spanning subgraph and robot odometer information.
Cartographer is cross-platform, Multi-sensor Fusion the laser SLAM algorithm of one kind that Google in 2016 proposes,
It is made of local optimum (Local SLAM) and global optimization (Global SLAM) two parts.
Attached drawing 2 is please referred to, in the Local Optimization Algorithm of front end, Cartographer is mainly to pass through what fitting was newly inserted into
Laser radar scanning point data and Inertial Measurement Unit (IMU) measurement acceleration and angle-data, will newly be inserted into data with it is right
The location matches problem between subgraph is answered to be equivalent to a least square problem, so that matched between scan data point set and subgraph
Maximum probability finds best insertion point of the data in subgraph with this, and after new number of scan points is according to insertion to subgraph into
Row real-time update.
The subgraph generated in Local Optimization Algorithm is the concept that Cartographer algorithm newly introduces, he is used to
It indicates a part in global map, and is presented in map in the form of probabilistic grid.Its each grid has one to consolidate
Fixed probability value indicates probability that the grid is blocked, can recalculate one in new number of scan points according to after being inserted into subgraph
It is a represent mesh point hits set and misses set, by with before subgraph hits set and misses set progress
The probability value p of representative mesh point is gathered hits and misses in comparison using new probability formula (2)hitsOr pmissIt carries out more
Newly.And current subgraph is exported after inserting enough scanning point sets, and start the update of new subgraph.
Mnew(x)=clamp (odd-1(odds(Mold(x))·odds(phits/misses))) (6)。
In the global optimization approach of rear end, Cartographer can by each subgraph that local optimization algorithm generates with
Number of scan points evidence and its position for forming it are recorded and are detected as winding, and fixed using branch in robot kinematics
The data being newly inserted into and the data comparison recorded before are established winding constraint by boundary's method windowing detection winding.Finally recycle
SPA (Sparse Pose Adjustment) algorithm optimizes the position of all subgraphs again.Fig. 6 (a) and Fig. 6 (b) be
It carries out building the effect picture before and after figure winding under real scene, it can be clearly seen that, it can be eliminated by winding amendment and largely be deposited
Accumulated error.
In the present embodiment, the robot odometer information is the routing information of robot, including machine in the unit time
Device people is in X-axis, the changes in coordinates of Y-axis and angle change.
S103 binds subgraph and corresponding points cloud that step 102 obtains, and splicing subgraph generates three-dimensional map.
In the present embodiment, by subgraph that Cartographer algorithm is generated as key frame, by corresponding cloud
What the mode being inserted into the corresponding space of subgraph was completed builds figure work.Simultaneously in rear end, Cartographer can to subgraph and
The scanning point set for being inserted into subgraph before carries out winding optimization, adjusts by adjusting the mode of subgraph posture information and is inserted into subgraph
Point cloud.So needing to tie up point cloud information and subgraph to enable the point cloud information between subgraph to be accurately superimposed
It is fixed, and find a kind of method that resource consumption is less and adjusted cloud is updated.
Specifically, subgraph and corresponding points cloud bound to specific step is as follows:
During robot advances, with the continuous renewal of laser radar scanning point, Cartographer can pass through ROS
The form of topic "/tf " constantly exports robot posture information.Recording the posture information is δ=[δx, δy, δz, δpitch,
δroll, δyaw]T, and depth image, the RGB image of the timestamp and real-time RGBD camera by comparison/tf topic and image topic
Carry out time synchronization.Meanwhile after a new subgraph generates, record the newly-generated subgraph posture information be φ=
[φx, φy, φz, φpitch, φroll, φyaw]TAnd timestamp, compared by the timestamp with image, selection near
The deep image information and robot posture information of subgraph.
Since Cartographer can be continuously generated subgraph in the process of running, with the superposition of subgraph, it is bundled in subgraph
The upper point Yun Yehui being registrated is overlapped therewith, that is, produces global map, and Fig. 5 is Submap during on-line operation14It ties up
Fixed point cloud and Submap15The effect of the point cloud superposition of binding.
During image information and robot posture information synchronize, the publication frequency of image information is 25HZ, machine
The publication frequency of device people's pose is 70HZ, and carrying out time synchronization between can control the time difference of the two in ± 14ms
It is interior.Different from above-mentioned synchronizing process, since the publication set of frequency of subgraph is 5HZ, Cartographer exports newly-generated son
When figure, the time error of received image data theory presence ± 40ms, but operation is more in actual moving process, especially
It is during winding, it is possible that program is blocked, the time error between subgraph and the point cloud data being matched to is general
Have 120ms or so.The posture δ and φ of point cloud received at this time are not necessarily equal, in the present embodiment, robot with
The speed of 2m/s is mobile, and generally there are the errors of 20cm or so between the robot pose that subgraph and subgraph are matched to.And
The attitude value of the newly-generated subgraph of Cartographer, which is directed toward default direction, represents the direction of the subgraph, in the scanning element that the later period is new
Data can constantly be adjusted this value after being added, and be different from the attitude value for being directed toward robot direction in robot pose
Two variables.Simultaneously Cartographer carry out winding optimization when the pose of winding point cloud will not be adjusted, but to return
The pose of the corresponding subgraph of circling point cloud is adjusted.
Therefore, it is necessary to convert to a cloud, the water between robot and subgraph is calculated by formula (7), formula (8)
Prosposition moves and the angle of rotation, and utilizes the Eigen library initialization spin matrix R of C++ΔTo Eulerian angles rΔDirection.It will rotation
Torque battle array RΔWith Eulerian angles tΔBring European transformation matrix Trans intoΔIn, it is right with it a cloud P can be transformed to using formula (9)
It answers in the referential of subgraph.
tΔ=[δx-φx, δy-φy, δz-φz]T (7)
rΔ=[δpitch-φpitch, δroll-φroll, δyaw-φyaw]T (8)
After the binding of the point cloud for transforming to subgraph referential and picture information and will record, according still further to formula (11) to cloud into
Line position is moved, and transformed cloud can be moved to Sub-graph Space position.
tSubmap=[φx, φy, φz]T (10)
Wherein, tSubmapFor the pose of corresponding subgraph in space；P is to be tied to the pose that the point in cloud is put on the subgraph
Information.
The present embodiment binds subgraph and corresponding points cloud, after winding, needs to be adjusted subgraph posture information
When, it does not need to carry out pose transformation to the point cloud being tied on the subgraph for needing to adjust pose again, shortens the time used in winding.
Global three-dimensional map is generated with subgraph is gradually spliced in Cartographer algorithm operational process, is bundled in correspondence
Point cloud on point Yun Yehui be spliced, with this come realize RGBD camera three-dimensional scene construct.
Is there is point cloud bound in the subgraph changed and pressed by S104, winding point cloud when subgraph pose changes in pose
It is translated according to subgraph variation.
With being continuously added for subgraph, Cartographer can carry out winding detection with the method for DFS branch boundary, and build
Vertical winding constrains, and adjusts the pose φ of subgraph in real time using SPA method.The pose φ for the subgraph established before i.e., not only can
It adjusts, the pose of subgraph can be also adjusted when carrying out the correction of sparse pose, subgraph pose tune when winding
Whole frequency can be relatively high.
In embodiment, a cloud is stored and is superimposed using the PointXYZRGB type in the library PCL, if every
The all the points cloud holding time stored before just translation superposition is primary again after secondary subgraph appearance variation is longer, cannot achieve reality
When point cloud superposition and display.
To solve this problem, when subgraph pose converts, subregion is converted and is superimposed.Use the class template in the library C++
Vector defines the region point cloud array of a PointXYZRGB type, and is inserted into after the point cloud of given threshold quantity is superimposed
Region point cloud array.When subgraph pose changes, only its corresponding region point cloud is translated and is superimposed again, and fixed
When device call back function in all areas point cloud be converted into ROS topic issue out.
Figure experiment is built, variation and again summing point cloud time-consuming 2.1s occurs when there are 11 subgraph poses, and do not having
The display frequency of RVIZ can stablize more than 5 frames (27 subgraph p), substantially to real-time display substantially in the case of point cloud is superimposed again
Do not influence.
Specifically, the specific implementation step of region point cloud superposition are as follows:
Obtain the subgraph of Catographer algorithm output, real-time cloud of generation；
Judge whether subgraph newly occur, if new occur, finds and time difference the smallest real-time cloud at this time；
The position that cloud is transformed to relatively new subgraph by formula (9) will be put in real time；
If the subgraph quantity in remnants point cloud is greater than after current emerging subgraph is added to newest remaining point cloud
The cloud is then added in remaining point cloud, and is recorded in memory as a region point cloud by given threshold, removes simultaneously
Content in remnants point cloud.
If the subgraph quantity in remnants point cloud does not reach after current emerging subgraph is added to newest remaining point cloud
To given threshold, then the cloud is added in remaining point cloud, is judged after waiting new subgraph appearance next time.
If it find that the subgraph pose generated before changes, judge the subgraph in which region point cloud, and it is right
Point cloud in the region point cloud is translated again according to formula (11), and is exported.
The large-scale three dimensional scene map that the present embodiment obtains can be used for the brainpower insufflation of green house of vegetables, and the present embodiment can be used
The RGBD camera three-dimensional scene construction method of proposition constructs green house of vegetables three-dimensional environment map, and controls robot in green house of vegetables
It is inside moved and plant is sprayed or applied fertilizer.
Experimental verification
The RGBD camera three-dimensional scene construction method based on Cartographer algorithm that the present embodiment proposes is carried out real
Verifying.In this experimentation, received using the Mike equipped with KinectV2 camera, SICKTIM561 laser radar and IMU
Nurse homotopy moving platform carries out field test.The computer for carrying out step in RGBD camera three-dimensional scene construction method is configured that
Macbookpro(Apple),i7 6700 2.6GHZCPU,16GRAM,512GRAM.During building figure, Kinect is selected to pass
Point cloud data in the higher section 0.2m to 5.5m of sensor depth data precision calculates, and uses the voxel of 0.03m sizing grid
Filter (Voxel Filter) is down-sampled to original point cloud and superimposed cloud progress, and by way of ROS topic
It sends point cloud data on RVIZ and shows.The selection of sample plot point is returned in the Stall hall of school district main building and second floor annular
Corridor, as shown in Fig. 7 (a), Fig. 7 (b), Fig. 7 (c), Fig. 7 (d), Fig. 7 (e) and Fig. 7 (f).And select accuracy and the speed of service
Opposite better RTAB (the Real Time of RGB-DSLAMV2 and DVOSLAM (Dense Visual Odometry and SLAM)
Appearance-Based Mapping) interior builds nomography and compares experiment.
First of all for rejection ability of the RGBD camera three-dimensional scene construction method to interference for verifying the present embodiment proposition, choosing
It selects and carries out field experiment in the even claustra hall of school district main building second floor uneven illumination, robot is at the uniform velocity run around claustra with 2m/s's
One circle half, with the help of claustra wall, the data volume that laser radar and RGBD camera measure is more, also can without winding amendment
Obtain preferable odometer precision.Second floor inlet is a face glass wall, has a large amount of sunlight and injects, by Fig. 7 (e) and Fig. 7
(f) depth image can be seen that the RGBD camera under the interference of sunlight and be lost most depth information, to using only vision
The RTAB algorithm of information is affected, and large error can be introduced in the calculating process of the visual odometry of RTAB, although counting
A cloud surveying range is not limited when calculation, is unable to complete three-dimensional map in repeatedly building in figure experimentation for progress
Building, all can lose robot location when close to doorway in each experimentation and cause to build figure failure.And only have
Once successfully to build three-dimensional map constructed in figure experiment also larger with reality scene gap, multiple places occur ghost image,
The phenomenon that dislocation and distortion, as shown in Fig. 7 (c) and Fig. 7 (d).And compared with RGBD camera, laser radar has stronger anti-noise
Performance, effectively measurement are built by the optimization function elimination of the winding of Cartographer most of during figure apart from longer
Accumulated error so that the present embodiment propose method advantage becomes apparent from three-dimensional reconstruction operation under large scene, can be big
More accurate three-dimensional map is constructed under scene, as shown in Fig. 7 (a) and Fig. 7 (b).
And in order to compare the precision that the method for the present embodiment is superimposed in large scene three-dimensionalreconstruction process midpoint cloud, night without
The sunlight and more uniform school district main building Stall of indoor light source has carried out comparative experiments, big chief 27m, width 16m, area is about
300m2, robot is with 2m/s at the uniform velocity in Stall greatly around one circle half of column operation.Figure effect is built as shown in Fig. 8 (c) and Fig. 8 (d)
It is preferably built as can be seen that being overlapped also obtain to point cloud data using the Submap pose that Cartographer is calculated
Figure effect, (such as Fig. 8 (e) and Fig. 8 compared with the resulting result of pose snap point cloud for using visual odometry in RTAB algorithm
(f) shown in), using this paper algorithm calculated cloud can be superimposed by more accurate alignment under the resolution ratio of 5cm
Point cloud substantially without ghost image.
Due to not being that can not set up optical tracking instrument in main building in laboratory environment verification algorithm.Therefore in order to verify
The method of the present embodiment and RTAB's builds figure precision, has carried out field survey to main building Stall, has first built to Cartographer
Figure precision is verified.As shown in figure 9, laterally the map distance error of building between two support columns is 2cm, longitudinal two support columns it
Between figure error of building be 1cm, the range error about 4cm of hall transverse direction longest part, the range error of longitudinal longest part is
7cm.In Real-Time Loop Closure in 2D LIDAR SLAM paper, Wolfgang Hess et al. experimental result
In error range be consistent.
Then the method for the present embodiment and RTAB algorithm track calculated are plotted on same figure, pass through two tracks
The distance between error build figure precision compare this paper algorithm and RTAB algorithm.By Figure 10 (a), Figure 10 (b), Figure 10 (c), figure
Shown in 10 (d).It can be seen that the RTAB algorithm of view-based access control model odometer is in open scene or meeting in the case where have a wide range of interference
There is the biggish calculating error greater than 1m, much larger than the error of Cartographer algorithm 5cm, therefore the method institute of the present embodiment
The three-dimensional map precision of building is higher.
The RGBD camera large-scale three dimensional scenario building method based on Cartographer algorithm that the present embodiment proposes, for
RGBD vision SLAM algorithm in the case where large scene and big interference is easily lost self-position and to build figure accuracy inadequate
Laser SLAM algorithm is applied and is built in figure in large scene three-dimensional by problem, and has carried out field experiment, online verification in school district main building
This paper algorithm has more real-time, accuracy and anti-dry with respect to the nomography of building that visual odometry data are used only in traditional sense
Disturb ability.
Embodiment two
The RGBD camera large-scale three dimensional scenario building system based on Cartographer algorithm that the present embodiment provides a kind of, should
System includes:
Point cloud registering module generates three-dimensional point cloud for obtaining the depth image and color image of RGBD camera, and by point
Cloud is registrated to laser radar referential；
Subgraph generation module is utilized for obtaining the number of scan points evidence of robot pose data and laser radar
The Local Optimization Algorithm of Cartographerd algorithm handles it, spanning subgraph and robot odometer information；
Point cloud binding module, for according to robot odometer information, the subgraph of generation to be tied up with corresponding cloud
It is fixed；
Three-dimensional building module generates three-dimensional map for splicing subgraph, and the point cloud being bundled on corresponding points cloud can also be spelled
Connect, with this come realize RGBD camera three-dimensional scene construct；
Winding point cloud module, for when subgraph pose changes, point bound in the subgraph changed to be occurred in pose
Cloud is translated again according to subgraph variation.
Embodiment three
A kind of computer readable storage medium is present embodiments provided, computer program is stored thereon with, which is located
It manages when device executes and realizes the RGBD camera large-scale three dimensional scenario building based on Cartographer algorithm as shown in Figure 1 or 2
Step in method.
Example IV
Present embodiments provide a kind of computer equipment, including memory, processor and storage are on a memory and can be
The computer program run on processor, the processor realized when executing described program it is as shown in Figure 1 or 2 based on
Step in the RGBD camera large-scale three dimensional scenario building method of Cartographer algorithm.
It should be understood by those skilled in the art that, embodiment of the disclosure can provide as method, system or computer program
Product.Therefore, the shape of hardware embodiment, software implementation or embodiment combining software and hardware aspects can be used in the disclosure
Formula.Moreover, the disclosure, which can be used, can use storage in the computer that one or more wherein includes computer usable program code
The form for the computer program product implemented on medium (including but not limited to magnetic disk storage and optical memory etc.).
The disclosure be referring to according to the method for the embodiment of the present invention, the process of equipment (system) and computer program product
Figure and/or block diagram describe.It should be understood that every one stream in flowchart and/or the block diagram can be realized by computer program instructions
The combination of process and/or box in journey and/or box and flowchart and/or the block diagram.It can provide these computer programs
Instruct the processor of general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to produce
A raw machine, so that being generated by the instruction that computer or the processor of other programmable data processing devices execute for real
The device for the function of being specified in present one or more flows of the flowchart and/or one or more blocks of the block diagram.
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.
Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, the program can be stored in a computer-readable storage medium
In, the program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein, the storage medium can be magnetic
Dish, CD, read-only memory (Read-Only Memory, ROM) or random access memory (Random
AccessMemory, RAM) etc..
Although above-mentioned be described in conjunction with specific embodiment of the attached drawing to the disclosure, model not is protected to the disclosure
The limitation enclosed, those skilled in the art should understand that, on the basis of the technical solution of the disclosure, those skilled in the art are not
Need to make the creative labor the various modifications or changes that can be made still within the protection scope of the disclosure.
Claims (10)
1. a kind of RGBD camera large-scale three dimensional scenario building method based on Cartographer algorithm, characterized in that this method
The following steps are included:
The depth image and color image of RGBD camera are obtained, generates three-dimensional point cloud, and point cloud registering to laser radar is referred to
System；
The number of scan points evidence for obtaining robot pose data and laser radar, utilizes the local optimum of Cartographerd algorithm
Algorithm handles it, spanning subgraph and robot odometer information；
According to robot odometer information, the subgraph of generation is bound with corresponding cloud, splicing subgraph generates dimensionally
Figure.
2. the RGBD camera large-scale three dimensional scenario building method according to claim 1 based on Cartographer algorithm,
It is characterized in that described cloud generation method are as follows:
RGBD camera is registrated, depth camera referential is registrated under color camera referential；
According to the three-dimensional position of any point in the corresponding calculated for pixel values color image in depth image after registration in space
Confidence breath；
Whole picture depth image is successively traversed, corresponding cloud of camera is obtained.
3. the RGBD camera large-scale three dimensional scenario building method according to claim 1 based on Cartographer algorithm,
It is characterized in that described include: by point cloud registering to the step of laser radar referential
Robot chassis is obtained to enamel the angle of deviation of the camera photocentre at a distance from laser radar center and its in yaw angle
Degree；
It is converted using coordinate of the transformation matrix to each point in camera point cloud, by the scanning point set of laser radar and puts cloud
It is registrated.
4. the RGBD camera large-scale three dimensional scenario building method according to claim 1 based on Cartographer algorithm,
It is characterized in that the subgraph by generation includes: the step of binding with a corresponding cloud
The posture information and timestamp of the subgraph generated are recorded, the timestamp generated when by spanning subgraph and the acquisition of RGBD camera are deep
The timestamp of degree image compares, and selects the RGBD camera depth image information on the time near subgraph；
Time synchronization will be carried out near the RGBD camera depth image information of subgraph and robot posture information on time；
The differential seat angle of horizontal displacement difference and rotation when calculating spanning subgraph between robot and subgraph, forms spin matrix, makes
In the referential that Cloud transform is corresponded to subgraph with spin matrix to it；
The point cloud transformed in subgraph referential and picture information are bound, and a cloud is displaced, by transformed cloud
Move to Sub-graph Space position.
5. the RGBD camera large-scale three dimensional scenario building method according to claim 1 based on Cartographer algorithm,
It is characterized in that further including that pose point cloud bound in the subgraph changed is occurred according to subgraph when subgraph pose changes
The step of variation is translated again.
6. the RGBD camera large-scale three dimensional scenario building method according to claim 5 based on Cartographer algorithm,
It is characterized in that described there is the step of point cloud bound in the subgraph changed is translated again according to subgraph variation packet for pose
It includes:
Judge the region point cloud where the subgraph changed occurs in pose, the point cloud in the region point cloud is changed again according to subgraph
Move to Sub-graph Space position.
7. a kind of RGBD camera large-scale three dimensional scenario building system based on Cartographer algorithm, characterized in that the system
Include:
Point cloud registering module generates three-dimensional point cloud, and point cloud is matched for obtaining the depth image and color image of RGBD camera
Standard arrives laser radar referential；
Subgraph generation module is utilized for obtaining the number of scan points evidence of robot pose data and laser radar
The Local Optimization Algorithm of Cartographerd algorithm handles it, spanning subgraph and robot odometer information；
Point cloud binding module, for according to robot odometer information, the subgraph of generation to be bound with corresponding cloud；
Three-dimensional building module generates three-dimensional map for splicing subgraph.
8. the RGBD camera large-scale three dimensional scenario building system according to claim 7 based on Cartographer algorithm,
It is characterized in that the system further include:
Winding point cloud module, for point cloud bound in the subgraph changed being occurred in pose and is pressed when subgraph pose changing
It is translated again according to subgraph variation.
9. a kind of computer readable storage medium, is stored thereon with computer program, characterized in that the program is executed by processor
The Shi Shixian RGBD camera large-scale three dimensional scene structure for example of any of claims 1-6 based on Cartographer algorithm
Step in construction method.
10. a kind of computer equipment including memory, processor and stores the meter that can be run on a memory and on a processor
Calculation machine program, characterized in that realize when the processor executes described program and be based on as of any of claims 1-6
Step in the RGBD camera large-scale three dimensional scenario building method of Cartographer algorithm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910452208.XA CN110163968B (en) | 2019-05-28 | 2019-05-28 | RGBD camera large three-dimensional scene construction method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910452208.XA CN110163968B (en) | 2019-05-28 | 2019-05-28 | RGBD camera large three-dimensional scene construction method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110163968A true CN110163968A (en) | 2019-08-23 |
| CN110163968B CN110163968B (en) | 2020-08-25 |
Family
ID=67629702
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910452208.XA Active CN110163968B (en) | 2019-05-28 | 2019-05-28 | RGBD camera large three-dimensional scene construction method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110163968B (en) |
Cited By (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110849352A (en) * | 2019-11-21 | 2020-02-28 | 大连理工大学 | Method for constructing grid map by fusing infrared camera, depth camera and binocular camera |
| CN110849351A (en) * | 2019-11-21 | 2020-02-28 | 大连理工大学 | Method for constructing grid map by using depth camera and binocular camera |
| CN111340834A (en) * | 2020-03-10 | 2020-06-26 | 山东大学 | Lining plate assembly system and method based on data fusion of laser radar and binocular camera |
| CN111429344A (en) * | 2020-02-19 | 2020-07-17 | 上海交通大学 | Laser S L AM closed loop detection method and system based on perceptual hashing |
| CN111461982A (en) * | 2020-03-30 | 2020-07-28 | 北京百度网讯科技有限公司 | Method and device for splicing point clouds |
| CN111757021A (en) * | 2020-07-06 | 2020-10-09 | 浙江大学 | Multi-sensor real-time fusion method for mobile robot remote takeover scene |
| CN111795687A (en) * | 2020-06-29 | 2020-10-20 | 深圳市优必选科技股份有限公司 | Robot map updating method and device, readable storage medium and robot |
| CN111915723A (en) * | 2020-08-14 | 2020-11-10 | 广东申义实业投资有限公司 | Indoor three-dimensional panorama construction method and system |
| CN112312113A (en) * | 2020-10-29 | 2021-02-02 | 贝壳技术有限公司 | Method, device and system for generating three-dimensional model |
| CN112462385A (en) * | 2020-10-21 | 2021-03-09 | 南开大学 | Map splicing and positioning method based on laser radar under outdoor large environment |
| CN112799095A (en) * | 2020-12-31 | 2021-05-14 | 深圳市普渡科技有限公司 | Static map generation method and device, computer equipment and storage medium |
| CN112894832A (en) * | 2019-11-19 | 2021-06-04 | 广东博智林机器人有限公司 | Three-dimensional modeling method, three-dimensional modeling device, electronic equipment and storage medium |
| CN113379815A (en) * | 2021-06-25 | 2021-09-10 | 中德（珠海）人工智能研究院有限公司 | Three-dimensional reconstruction method and device based on RGB camera and laser sensor and server |
| WO2021189194A1 (en) * | 2020-03-23 | 2021-09-30 | 罗伯特博世有限公司 | Three-dimensional environment modeling method and device, computer storage medium, and industrial robot operating platform |
| WO2021208442A1 (en) * | 2020-04-14 | 2021-10-21 | 广东博智林机器人有限公司 | Three-dimensional scene reconstruction system and method, device, and storage medium |
| CN113587933A (en) * | 2021-07-29 | 2021-11-02 | 山东山速机器人科技有限公司 | Indoor mobile robot positioning method based on branch-and-bound algorithm |
| CN113674399A (en) * | 2021-08-16 | 2021-11-19 | 杭州图灵视频科技有限公司 | Mobile robot indoor three-dimensional point cloud map construction method and system |
| CN113834479A (en) * | 2021-09-03 | 2021-12-24 | Oppo广东移动通信有限公司 | Map generation method, device, system, storage medium and electronic equipment |
| CN113870358A (en) * | 2021-09-17 | 2021-12-31 | 聚好看科技股份有限公司 | Method and equipment for joint calibration of multiple 3D cameras |
| CN114063099A (en) * | 2021-11-10 | 2022-02-18 | 厦门大学 | RGBD-based positioning method and device |
| WO2022257801A1 (en) * | 2021-06-09 | 2022-12-15 | 山东大学 | Slam-based mobile robot mine scene reconstruction method and system |
| US11928830B2 (en) | 2021-12-22 | 2024-03-12 | Honeywell International Inc. | Systems and methods for generating three-dimensional reconstructions of environments |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2302586A1 (en) * | 2009-08-21 | 2011-03-30 | Sony Corporation | Information processing device, information processing method and program |
| US8224097B2 (en) * | 2008-06-12 | 2012-07-17 | Sri International | Building segmentation for densely built urban regions using aerial LIDAR data |
| US9390344B2 (en) * | 2014-01-09 | 2016-07-12 | Qualcomm Incorporated | Sensor-based camera motion detection for unconstrained slam |
| CN105783913A (en) * | 2016-03-08 | 2016-07-20 | 中山大学 | SLAM device integrating multiple vehicle-mounted sensors and control method of device |
| CN105843223A (en) * | 2016-03-23 | 2016-08-10 | 东南大学 | Mobile robot three-dimensional mapping and obstacle avoidance method based on space bag of words model |
| CN106595659A (en) * | 2016-11-03 | 2017-04-26 | 南京航空航天大学 | Map merging method of unmanned aerial vehicle visual SLAM under city complex environment |
| CN106598052A (en) * | 2016-12-14 | 2017-04-26 | 南京阿凡达机器人科技有限公司 | Robot security inspection method based on environment map and robot thereof |
| CN106910217A (en) * | 2017-03-17 | 2017-06-30 | 驭势科技（北京）有限公司 | Vision map method for building up, computing device, computer-readable storage medium and intelligent vehicle |
| CN107301654A (en) * | 2017-06-12 | 2017-10-27 | 西北工业大学 | A kind of positioning immediately of the high accuracy of multisensor is with building drawing method |
| US20180074176A1 (en) * | 2016-09-14 | 2018-03-15 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Motion compensation method and apparatus applicable to laser point cloud data |
| CN109507677A (en) * | 2018-11-05 | 2019-03-22 | 浙江工业大学 | A kind of SLAM method of combination GPS and radar odometer |
| CN109541630A (en) * | 2018-11-22 | 2019-03-29 | 武汉科技大学 | A method of it is surveyed and drawn suitable for Indoor environment plane 2D SLAM |
| CN109633664A (en) * | 2018-12-29 | 2019-04-16 | 南京理工大学工程技术研究院有限公司 | Joint positioning method based on RGB-D Yu laser odometer |
-
2019
- 2019-05-28 CN CN201910452208.XA patent/CN110163968B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8224097B2 (en) * | 2008-06-12 | 2012-07-17 | Sri International | Building segmentation for densely built urban regions using aerial LIDAR data |
| EP2302586A1 (en) * | 2009-08-21 | 2011-03-30 | Sony Corporation | Information processing device, information processing method and program |
| US9390344B2 (en) * | 2014-01-09 | 2016-07-12 | Qualcomm Incorporated | Sensor-based camera motion detection for unconstrained slam |
| CN105783913A (en) * | 2016-03-08 | 2016-07-20 | 中山大学 | SLAM device integrating multiple vehicle-mounted sensors and control method of device |
| CN105843223A (en) * | 2016-03-23 | 2016-08-10 | 东南大学 | Mobile robot three-dimensional mapping and obstacle avoidance method based on space bag of words model |
| US20180074176A1 (en) * | 2016-09-14 | 2018-03-15 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Motion compensation method and apparatus applicable to laser point cloud data |
| CN106595659A (en) * | 2016-11-03 | 2017-04-26 | 南京航空航天大学 | Map merging method of unmanned aerial vehicle visual SLAM under city complex environment |
| CN106598052A (en) * | 2016-12-14 | 2017-04-26 | 南京阿凡达机器人科技有限公司 | Robot security inspection method based on environment map and robot thereof |
| CN106910217A (en) * | 2017-03-17 | 2017-06-30 | 驭势科技（北京）有限公司 | Vision map method for building up, computing device, computer-readable storage medium and intelligent vehicle |
| CN107301654A (en) * | 2017-06-12 | 2017-10-27 | 西北工业大学 | A kind of positioning immediately of the high accuracy of multisensor is with building drawing method |
| CN109507677A (en) * | 2018-11-05 | 2019-03-22 | 浙江工业大学 | A kind of SLAM method of combination GPS and radar odometer |
| CN109541630A (en) * | 2018-11-22 | 2019-03-29 | 武汉科技大学 | A method of it is surveyed and drawn suitable for Indoor environment plane 2D SLAM |
| CN109633664A (en) * | 2018-12-29 | 2019-04-16 | 南京理工大学工程技术研究院有限公司 | Joint positioning method based on RGB-D Yu laser odometer |
Non-Patent Citations (2)
| Title |
|---|
| YINFA YAN 等: "Design and evaluation of visual SLAM method based on Realsense for mobile robots", 《2018 CHINESE AUTOMATION CONGRESS (CAC)》 * |
| 陈世浪 等: "基于RGB-D相机的SLAM技术研究综述", 《计算机是工程与应用》 * |
Cited By (32)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112894832A (en) * | 2019-11-19 | 2021-06-04 | 广东博智林机器人有限公司 | Three-dimensional modeling method, three-dimensional modeling device, electronic equipment and storage medium |
| CN112894832B (en) * | 2019-11-19 | 2022-06-03 | 广东博智林机器人有限公司 | Three-dimensional modeling method, three-dimensional modeling device, electronic equipment and storage medium |
| CN110849351A (en) * | 2019-11-21 | 2020-02-28 | 大连理工大学 | Method for constructing grid map by using depth camera and binocular camera |
| CN110849352A (en) * | 2019-11-21 | 2020-02-28 | 大连理工大学 | Method for constructing grid map by fusing infrared camera, depth camera and binocular camera |
| CN111429344A (en) * | 2020-02-19 | 2020-07-17 | 上海交通大学 | Laser S L AM closed loop detection method and system based on perceptual hashing |
| CN111429344B (en) * | 2020-02-19 | 2022-04-26 | 上海交通大学 | Laser SLAM closed loop detection method and system based on perceptual hashing |
| CN111340834A (en) * | 2020-03-10 | 2020-06-26 | 山东大学 | Lining plate assembly system and method based on data fusion of laser radar and binocular camera |
| CN111340834B (en) * | 2020-03-10 | 2023-05-12 | 山东大学 | Lining plate assembly system and method based on laser radar and binocular camera data fusion |
| WO2021189194A1 (en) * | 2020-03-23 | 2021-09-30 | 罗伯特博世有限公司 | Three-dimensional environment modeling method and device, computer storage medium, and industrial robot operating platform |
| CN111461982A (en) * | 2020-03-30 | 2020-07-28 | 北京百度网讯科技有限公司 | Method and device for splicing point clouds |
| CN111461982B (en) * | 2020-03-30 | 2023-09-22 | 北京百度网讯科技有限公司 | Method and apparatus for splice point cloud |
| CN113592989B (en) * | 2020-04-14 | 2024-02-20 | 广东博智林机器人有限公司 | Three-dimensional scene reconstruction system, method, equipment and storage medium |
| WO2021208442A1 (en) * | 2020-04-14 | 2021-10-21 | 广东博智林机器人有限公司 | Three-dimensional scene reconstruction system and method, device, and storage medium |
| CN113592989A (en) * | 2020-04-14 | 2021-11-02 | 广东博智林机器人有限公司 | Three-dimensional scene reconstruction system, method, equipment and storage medium |
| CN111795687A (en) * | 2020-06-29 | 2020-10-20 | 深圳市优必选科技股份有限公司 | Robot map updating method and device, readable storage medium and robot |
| CN111757021B (en) * | 2020-07-06 | 2021-07-20 | 浙江大学 | Multi-sensor real-time fusion method for mobile robot remote takeover scene |
| CN111757021A (en) * | 2020-07-06 | 2020-10-09 | 浙江大学 | Multi-sensor real-time fusion method for mobile robot remote takeover scene |
| CN111915723A (en) * | 2020-08-14 | 2020-11-10 | 广东申义实业投资有限公司 | Indoor three-dimensional panorama construction method and system |
| CN112462385A (en) * | 2020-10-21 | 2021-03-09 | 南开大学 | Map splicing and positioning method based on laser radar under outdoor large environment |
| CN112312113A (en) * | 2020-10-29 | 2021-02-02 | 贝壳技术有限公司 | Method, device and system for generating three-dimensional model |
| CN112799095B (en) * | 2020-12-31 | 2023-03-14 | 深圳市普渡科技有限公司 | Static map generation method and device, computer equipment and storage medium |
| CN112799095A (en) * | 2020-12-31 | 2021-05-14 | 深圳市普渡科技有限公司 | Static map generation method and device, computer equipment and storage medium |
| WO2022257801A1 (en) * | 2021-06-09 | 2022-12-15 | 山东大学 | Slam-based mobile robot mine scene reconstruction method and system |
| CN113379815A (en) * | 2021-06-25 | 2021-09-10 | 中德（珠海）人工智能研究院有限公司 | Three-dimensional reconstruction method and device based on RGB camera and laser sensor and server |
| CN113587933A (en) * | 2021-07-29 | 2021-11-02 | 山东山速机器人科技有限公司 | Indoor mobile robot positioning method based on branch-and-bound algorithm |
| CN113587933B (en) * | 2021-07-29 | 2024-02-02 | 山东山速机器人科技有限公司 | Indoor mobile robot positioning method based on branch-and-bound algorithm |
| CN113674399A (en) * | 2021-08-16 | 2021-11-19 | 杭州图灵视频科技有限公司 | Mobile robot indoor three-dimensional point cloud map construction method and system |
| CN113834479A (en) * | 2021-09-03 | 2021-12-24 | Oppo广东移动通信有限公司 | Map generation method, device, system, storage medium and electronic equipment |
| CN113870358A (en) * | 2021-09-17 | 2021-12-31 | 聚好看科技股份有限公司 | Method and equipment for joint calibration of multiple 3D cameras |
| CN113870358B (en) * | 2021-09-17 | 2024-05-24 | 聚好看科技股份有限公司 | Method and equipment for jointly calibrating multiple 3D cameras |
| CN114063099A (en) * | 2021-11-10 | 2022-02-18 | 厦门大学 | RGBD-based positioning method and device |
| US11928830B2 (en) | 2021-12-22 | 2024-03-12 | Honeywell International Inc. | Systems and methods for generating three-dimensional reconstructions of environments |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110163968B (en) | 2020-08-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110163968A (en) | RGBD camera large-scale three dimensional scenario building method and system | |
| CN104330074B (en) | Intelligent surveying and mapping platform and realizing method thereof | |
| CN113870343B (en) | Relative pose calibration method, device, computer equipment and storage medium | |
| CN102609942B (en) | Depth map is used to carry out mobile camera location | |
| Xie et al. | Study on construction of 3D building based on UAV images | |
| CN103226838A (en) | Real-time spatial positioning method for mobile monitoring target in geographical scene | |
| Xiao et al. | 3D point cloud registration based on planar surfaces | |
| CN107036594A (en) | The positioning of intelligent Power Station inspection intelligent body and many granularity environment perception technologies | |
| CN114419147A (en) | Rescue robot intelligent remote human-computer interaction control method and system | |
| CN115388902B (en) | Indoor positioning method and system, AR indoor positioning navigation method and system | |
| Kühner et al. | Large-scale volumetric scene reconstruction using lidar | |
| Hu et al. | Stable least-squares matching for oblique images using bound constrained optimization and a robust loss function | |
| Ceriani et al. | Pose interpolation slam for large maps using moving 3d sensors | |
| CN110260866A (en) | A kind of robot localization and barrier-avoiding method of view-based access control model sensor | |
| CN104318605A (en) | Parallel lamination rendering method of vector solid line and three-dimensional terrain | |
| CN118314300B (en) | Engineering measurement accurate positioning and three-dimensional modeling method and system | |
| Schadler et al. | Multi-resolution surfel mapping and real-time pose tracking using a continuously rotating 2D laser scanner | |
| CN114663473A (en) | Personnel target positioning and tracking method and system based on multi-view information fusion | |
| Chen et al. | Low cost and efficient 3D indoor mapping using multiple consumer RGB-D cameras | |
| CN117518196A (en) | Motion compensation method, device, system, equipment and medium for laser radar | |
| CN117007065B (en) | Unmanned aerial vehicle route planning method for planar slope wall-mounted flight | |
| Ma et al. | Low‐Altitude Photogrammetry and Remote Sensing in UAV for Improving Mapping Accuracy | |
| Xu et al. | A flexible 3D point reconstruction with homologous laser point array and monocular vision | |
| CN106959101A (en) | A kind of indoor orientation method based on optical flow method | |
| Wang et al. | Ct-LVI: A Framework Towards Continuous-time Laser-Visual-Inertial Odometry and Mapping |
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 | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20230105 Address after: 1201-7, Floor 12, Hanyu Jingu Artificial Intelligence Building, Jingshi Road, Jinan Area, China (Shandong) Pilot Free Trade Zone, 250000 Shandong Province Patentee after: Shandong Xinchen Artificial Intelligence Technology Co.,Ltd. Address before: 250061, No. ten, No. 17923, Lixia District, Ji'nan City, Shandong Province Patentee before: SHANDONG University |