CN109991636A - Map constructing method and system based on GPS, IMU and binocular vision - Google Patents
Map constructing method and system based on GPS, IMU and binocular vision Download PDFInfo
- Publication number
- CN109991636A CN109991636A CN201910225839.8A CN201910225839A CN109991636A CN 109991636 A CN109991636 A CN 109991636A CN 201910225839 A CN201910225839 A CN 201910225839A CN 109991636 A CN109991636 A CN 109991636A
- Authority
- CN
- China
- Prior art keywords
- imu
- coordinate system
- acceleration
- vision
- moment
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
- G01C21/28—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network with correlation of data from several navigational instruments
- G01C21/30—Map- or contour-matching
- G01C21/32—Structuring or formatting of map data
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S19/00—Satellite radio beacon positioning systems; Determining position, velocity or attitude using signals transmitted by such systems
- G01S19/38—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system
- G01S19/39—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system the satellite radio beacon positioning system transmitting time-stamped messages, e.g. GPS [Global Positioning System], GLONASS [Global Orbiting Navigation Satellite System] or GALILEO
- G01S19/42—Determining position
- G01S19/45—Determining position by combining measurements of signals from the satellite radio beacon positioning system with a supplementary measurement
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S19/00—Satellite radio beacon positioning systems; Determining position, velocity or attitude using signals transmitted by such systems
- G01S19/38—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system
- G01S19/39—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system the satellite radio beacon positioning system transmitting time-stamped messages, e.g. GPS [Global Positioning System], GLONASS [Global Orbiting Navigation Satellite System] or GALILEO
- G01S19/42—Determining position
- G01S19/45—Determining position by combining measurements of signals from the satellite radio beacon positioning system with a supplementary measurement
- G01S19/47—Determining position by combining measurements of signals from the satellite radio beacon positioning system with a supplementary measurement the supplementary measurement being an inertial measurement, e.g. tightly coupled inertial
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Networks & Wireless Communication (AREA)
- Automation & Control Theory (AREA)
- Navigation (AREA)
Abstract
The present invention discloses a kind of map constructing method and system based on GPS, IMU and binocular vision, is corrected using IMU and binocular vision to it and uses vision guided navigation without the weak place of GPS signal or signal in tunnel etc..It can be good at solving the problems in vision SLAM, and improve the robustness of SLAM system.The method merged using visual information and IMU information, multiple sensors data are merged, provide new algorithm structure, effective frame is stablized in building, pilotless automobile or robot are carried out stablizing effective autonomous positioning and map structuring, it realizes that the location information of GPS is corrected and supplements, is applied to pilotless automobile and robot system, carry out large-scale application.
Description
Technical field:
The present invention discloses a kind of map constructing method and system based on GPS, IMU and binocular vision, is related to one kind
The autonomous positioning of view-based access control model feature and IMU information and map structuring system, belong to robot and control technology field.
Background technique:
In recent years, the rapid development of Internet technology brings the chance of revolutionary variation to automobile industry.With this
Meanwhile vehicle intellectualized technology is just gradually used widely, this technology simplifies the driver behavior of automobile and improves row
Sail safety.And it is exactly pilotless automobile that wherein most typically, which is also the most popular following application,.In adding for artificial intelligence technology
It holds down, unmanned high speed development, is changing the trip mode of the mankind, and then relevant industries pattern can be changed on a large scale.
For travelling pilotless automobile in zone of ignorance, due to being blocked etc. building, trees, automobile is normal
The often state in no signal or weak signal, can not provide effective position；In the place of bad environments, because of the reasons such as weather GPS
Or Beidou Navigation System weak output signal, pilotless automobile can not effectively be positioned.For this purpose, pilotless automobile is necessary
Ability with autonomous positioning and map structuring.By real-time autonomous positioning and map structuring, ambient condition information is obtained, really
Determine the location of pilotless automobile, important foundation is provided for path planning.
Simultaneous localization and mapping (simultaneous localization an mapping, SLAM) technology is machine
The important problem in one, device people field, SLAM problem can be described as: the robot for being equipped with sensor freely moves in space
It is dynamic, using collected sensor information, self-position is positioned, while on the basis of self poisoning increment type structure
Map is built, positions and charts while realizing robot.Two key factors for influencing SLAM Resolving probiems are respectively sensor
Data characteristics and observation data dependence, if it is possible to it is efficient using sensing data and to improve the accuracy of data correlation
And robustness, the positioning of robot and the building of map will be influenced.
However unmanned technology of today is mostly still in auxiliary driving phase, lacks autonomous positioning and map structuring
Ability.Pilotless automobile less simultaneously carries out autonomous positioning and map structuring, positioning and map structure using single-sensor
It is lower to build precision, cannot effectively merge multiple sensors, or multiple sensors data can be merged, but
It is that cannot carry out stablizing effective autonomous positioning and map structuring to pilotless automobile, and cannot popularize on a large scale.
Summary of the invention:
The present invention provides a kind of map constructing method and system based on GPS, IMU and binocular vision, using IMU and
Binocular vision is corrected it and uses vision guided navigation without the weak place of GPS signal or signal in tunnel etc..It can be good at
The problems in vision SLAM is solved, and improves the robustness of SLAM system.
A kind of map constructing method and system based on GPS, IMU and binocular vision of the present invention, the skill of use
Art scheme are as follows:
A kind of map constructing method based on GPS, IMU and binocular vision of the present invention, which is characterized in that including
Following steps:
Step 1: the pose of initialization pilotless automobile or robot, determines pilotless automobile traveling or robot
The track of movement corrects IMU Inertial Measurement Unit, the first-class data hold time consistency of binocular camera shooting；
Step 2: carrying out IMU Inertial Measurement Unit and binocular camera and the letter based on the RTK GNSS receiver resolved
Breath acquisition and extraction, to carry out visual signature tracking；
Step 3: the IMU pre-integration based on flow pattern, comprising: IMU pre-integration calculates, calculates pre- score Jacobi and association side
Poor matrix calculates residual error Jacobi；
Step 4: IMU initialization (initialization of vision inertia joint), this is the first step of vision and inertial data fusion,
It is one section of loose coupling process, vision data and quick IMU data is combined, comprising: offset of gyroscope demarcates (zero bias), ruler
Degree restores and acceleration of gravity pre-estimation, accelerometer biasing calibration and the optimization of scale acceleration of gravity；
Step 5: being optimized using close coupling Optimized model, comprising: the optimization of vision inertia consecutive frame close coupling, vision
The optimization of inertia local map close coupling；
Step 6: estimation pilotless automobile or the similarity degree of robot trajectory and ambient enviroment three-dimensional map, obtain
Real-time pose；
Step 7: constructing map and carrying out closed loop detection using bag of words, existing map is corrected, is further mentioned
The precision of high three-dimensional map.
1, in step 2) information collection and extracting:
After each binocular camera collects image, the extraction of ORB feature is carried out to image, it, will after extracting ORB feature
The result of extraction stores；
The IMU data of pilotless automobile or robot interior are acquired in real time, and it is pre- to be sent into the IMU based on flow pattern
It is handled in integral.
2. in IMU pre-integration of the step 3) based on flow pattern:
During the IMU pre-integration calculates, world coordinate system, IMU coordinate system and camera coordinates system usually with W, B, C come
It indicates, Δ t is sampled at a certain time interval, measures the acceleration a of IMUBAnd angular velocity omegaB.IMU measurement model:
Wherein,The angular speed and acceleration of IMU measurement model under indicates coordinate system B,BωWB(t)∈R3It indicates
Instantaneous angular velocity of the center the IMU B relative to world coordinate system W, R at coordinate system BWBIndicate the center IMU B relative to world coordinates
It is the spin matrix of W,For RWBTransposition,Wa(t)∈R3Indicate the instantaneous acceleration of the IMU under W coordinate,WG indicates to sit in W
Mark is lower gravitational vectors, and b indicates zero bias, and η indicates gaussian random noise, ba、bgAcceleration and gravitational vectors zero bias are respectively indicated,
ηa、ηgRespectively indicate acceleration and gravitational vectors gaussian random noise；
Rotation R, the derivative for translating p and speed v can indicate are as follows:
Wherein,WV indicates the instantaneous velocity of the IMU under W coordinate system,Indicate instantaneous angular velocity of the B relative to W, RWBTable
Show spin matrix of the center the IMU B relative to world coordinate system W,WP indicates the instantaneous translation under W coordinate system,
For its derivative；
Rotation R, translation p and speed v under world coordinate system can be acquired by general integral formula, be used under discrete time
Euler's integral can will above continuous time integral rewrite and simultaneous can obtain:
Δ t is certain time interval, ηgd(t)、ηad(t) indicate under discrete time acceleration and gravitational vectors Gauss with
Machine noise, ba、bgRespectively indicate acceleration and gravitational vectors zero bias, ηad、ηgdRespectively indicate acceleration and gravitational vectors Gauss with
Machine noise,For the angular speed and acceleration of IMU measurement model, quantity of state at this time is under world coordinate system；
If needing corresponding IMU quantity of state of known i moment now with two adjacent key frames of i and j, discrete time is solved
The corresponding IMU pre-integration value of lower j moment key frame:
Δ t is certain time interval, Δ tijFor the time difference between i, j moment, g is gravitational vectors,Respectively
Indicate the gravitational vectors changed over time and acceleration zero bias,Indicate the IMU measurement model angular speed changed over time,
Indicate the IMU measurement model speed changed over time, Rj、Ri、vj、vi、pj、piThe respectively rotation at j moment and i moment, speed
And translation, RkFor the spin matrix changed over time,For discrete acceleration and the gravitational vectors height changed over time
This random noise；
In order to avoid repetition integral is changed using the opposite of quantity of state that IMU pre-integration method calculates i, j moment, that is,
Reference frame is transformed under the i moment coordinate system of IMU by world coordinate system, IMU quantity of state is calculated at i the and j moment
Between relative quantity:
Rj、Ri、vj、vi、pj、piThe respectively rotation at j moment and i moment, speed and translation,Point
The acceleration and gravitational vectors zero bias that Wei do not change over time,Indicate the IMU measurement model angle changed over time
Speed,Indicate the IMU measurement model speed changed over time,For the discrete acceleration and gravity changed over time
Vector Gaussian random noise, Δ Rij、Δvij、ΔpijRotation, speed and translation between i, j is poor,For Δ RiTransposition,
ΔtijFor the time difference between i, j moment, Δ t is certain time interval, Δ Rik、ΔvikFor the rotation and speed at i, k moment
Difference；
From the equations above as can be seen that entire pre-integration formula and biasing, noise have relationship, need to analyze respectively partially
The influence with noise to system is set, noise obeys Gauss normal distribution, bigoted obedience random walk model；
It is the single order item not biased in the IMU pre-integration that initialization uses, Jacobian matrix is in initialization completion
It is just calculated before optimizing afterwards, obtains the accurate pre-integration value (true value) of IMU and estimated value to obtain IMU's
Residual error；
System is Gaussian distributed, and covariance matrix is calculated according to Gaussian Profile, is the matrix of a 9*9；
The calculating residual error Jacobi:
Residual error:
Parameter to be estimated:
ΔRij、Δvij、ΔpijIt is poor for the rotation at i, j moment, speed and translation,For device average value, For the acceleration and gravitational vectors zero bias at i moment,For its average value, δ ba, δ bg are its renewal amount, and φ is rotation
Turn, Δ tijInter frame temporal is poor, Ri、Rj、vi、vj、pi、pjIndicate spin matrix, speed and the translation at i, j moment,For its turn
It sets；
The corresponding Jacobi of displacement residual error is as above, and the j moment, corresponding Jacobi can also so derive, the rotation of speed residual sum
Jacobi's formula it is similar to translation.So far IMU pre-integration is completed.
3. in step 4) IMU initialization (initialization of vision inertia joint):
The rotation amount of key frame is contacted using two to estimate the biasing of gyroscope, in a constant bgOn the basis of add
One smallChange, calculate the renewal amount of offset of gyroscope by minimizing rotation error:
Wherein, N is the number of key frame, Δ Ri,i+1It is two continuous key frame pre-integration wheel measuring amounts, J is covariance
Matrix,Respectively the B coordinate system at i+1 moment is relative to W coordinate system and the W coordinate system at i moment relative to B coordinate
The spin matrix of system；Give zero initial value, above-mentioned function is optimized by G2O, and it is corresponding to solve minimum target functionAs top
The initial bias of spiral shell instrument；
Pre-integration formula is substituted into after finding out zero bias can recalculate a pre-integration value, make pre-integration numerical value more
Accurately；
It carries out scale to restore and acceleration of gravity pre-estimation, it is necessary first to establish pre-evaluation state vector X=[s, gW],
Wherein, s is scale, gWIt is the acceleration of gravity under world coordinate system；
Three key frames are used herein (indicates) that simultaneous vision and IMU pre-integration data construct an AX with 1,2,3
The least square overdetermined equation of=B at least needs four key frames, seeks least square problem using singular value decomposition, speed compared with
Slow but precision is high；
I is unit matrix,WPCFor the displacement of image center C under W coordinate system, RWB、RWCIt is W coordinate system relative to B coordinate system
With the spin matrix of C coordinate system, Δ tnmTime difference (n, m=1,2,3) between two frames；
Since scale restores the influence for not accounting for accelerometer biasing with the calculating process of acceleration of gravity pre-estimation,
So that acceleration of gravity and accelerometer biasing are difficult to differentiate between, it is likely that lead to systemic disease state property problem, so carrying out acceleration
Meter biasing calibration and the optimization of scale acceleration of gravity；
If the acceleration of gravity direction of inertial coodinate system IgWNormalizated unit vectorθ is angle.Then the spin matrix between world coordinate system W and inertial coodinate system I is calculated:
Then acceleration of gravity is expressed as:
Due to RWIIt is rotation about the z axis, so only by X, the influence of Y-direction:
Biasing comprising acceleration:
Using three continuous key frames come release rate amount:
Wherein [](:,1:2)The first two columns of representing matrix, acceleration bias renewal amountδ θ is minute angle variable,WPCFor
The displacement of image center C under W coordinate system,CPBFor translation of the center the IMU B under C coordinate system, s is scale, Δ tnm、Δpnm、Δ
vnmIt is time, displacement and the speed difference (n, m=1,2,3) between two frames, R respectivelyWB、RWC、RWIIt is W coordinate system relative to B coordinate
It is, the spin matrix of C coordinate system and I coordinate system,For g under I coordinate systemINormalizated unit vector, J is covariance matrix；
Only optimize the direction of acceleration of gravity, the data that every suboptimization obtains all can calculate newest g by substituting into againWContinue excellent
Change, variable convergence of going directly.
In step 5) is optimized using close coupling Optimized model:
There are two kinds of optimal ways for the optimization of vision inertia consecutive frame close coupling: (1) there is map rejuvenation:
Global optimization state vector, the rotation comprising current time j are constructed firstTranslational velocityDisplacement
Accelerometer biasingAnd offset of gyroscopeGlobal error equation includes vision re-projection error EprojWith IMU measurement error
EIMU；
Vision re-projection equation is exactly simple pinhole camera re-projection equation, there is no what especially its residual error, it is refined can
It is no different than calculating and information matrix with general vision；
IMU measuring error equation includes two parts, is P, V, R and b respectivelya,bgBiasing；Its residual error, residual error Jacobian matrix
Completion has been settled accounts in IMU pre-integration process with information matrix, has been brought directly to:
eb=bj-bj
(40)
eR、ep、ev、ebFor rotation, translation, speed and biased error, RWB、RWC、RWIFor W coordinate system relative to B coordinate system,
The spin matrix of C coordinate system and I coordinate system, b are biasing；
(2) there is not map rejuvenation:
As described above, constructing global optimization state vector, the rotation including current time j+1 and last moment j first Translational velocityDisplacementAccelerometer biasingAnd offset of gyroscopeGlobal error equation includes vision re-projection error Eproj, IMU measurement error EIMUWith prior uncertainty Eprior:
The error equation of vision and IMU and as before, prior uncertainty is the quantity of state at j moment；
eR、ep、ev、ebFor rotation, translation, speed and biased error,For its transposition, RWBFor W seat
Mark system is relative to B coordinate system spin matrix, RBWIt is B coordinate system relative to W coordinate system spin matrix, b is biasing, and ρ is coefficient,WpB、WvBThe translation and speed that are the center IMU B under W coordinate system；
Vision inertia local map close coupling optimize when, the key frame for needing to optimize include vision re-projection error and
IMU measurement error.
Step 7) closed loop detection in, building map simultaneously using bag of words progress closed loop detection, to existing map into
Row correction, further increases the precision of three-dimensional map:
A kind of map structuring system based on GPS, IMU and binocular vision of the present invention, which includes vehicle-mounted
Arithmetic center, binocular camera, IMU Inertial Measurement Unit, the GNSS receiver based on RTK resolving are electrically connected；
The vehicle-mounted arithmetic center is by ROS frame by binocular camera, IMU Inertial Measurement Unit and based on RTK resolving
The collected data of GNSS receiver are stored and are handled；
The GNSS receiver resolved based on RTK receives GPS positioning information；
The binocular camera acquisition surrounding image information carries out point Yun Jianmo；
The IMU Inertial Measurement Unit acquisition measurement automobile triaxial attitude angle and acceleration；
The pose for initializing pilotless automobile or robot first determines pilotless automobile traveling or robot motion
Track, correct IMU Inertial Measurement Unit, the first-class data hold time consistency of binocular camera shooting；IMU inertia measurement is carried out later
The information collection and extraction of unit and binocular camera and the GNSS receiver based on RTK resolving, are sent into vehicle-mounted arithmetic center；
Secondly the IMU pre-integration based on flow pattern is carried out in vehicle-mounted arithmetic center, IMU initialization (initialization of vision inertia joint), is made
Optimized with close coupling Optimized model and is calculated；Estimate the similar of pilotless automobile or robot trajectory and ambient enviroment three-dimensional map
Degree obtains real-time pose；It constructs map and carries out closed loop detection using bag of words, existing map is corrected, into one
Step improves the precision of three-dimensional map.
The positive effect of the present invention is:
Using the method that visual information and IMU information merge, multiple sensors data are merged, are provided new
Effective frame is stablized in algorithm structure, building, carries out stablizing effective autonomous positioning and ground to pilotless automobile or robot
Figure building, realizes that the location information of GPS is corrected and supplements, and is applied to pilotless automobile and robot system, carries out big
Sizable application.
Detailed description of the invention:
Fig. 1 is system structure diagram of the invention；
Fig. 2, Fig. 3 are picture captured by binocular camera when embodiment 2 is tested in Chengdu University of Electronic Science and Technology campus.
Specific implementation method:
Below in conjunction with attached drawing, preferred implementation of the invention is described in detail.It should be appreciated that preferred implementation is only
Illustrate the present invention, rather than limiting the scope of protection of the present invention.
Embodiment 1
The present invention is based on the autonomous positioning and map constructing method of GPS, IMU and binocular vision and system, including it is following
Step:
1, the pose for initializing pilotless automobile or robot determines pilotless automobile traveling or robot motion
Track, the data hold time consistency of correction IMU Inertial Measurement Unit, binocular camera, guarantees it under same local area network；
2, IMU Inertial Measurement Unit and binocular camera are carried out and the information based on the RTK GNSS receiver resolved is adopted
Collection and extraction, to carry out visual signature tracking；
3, the IMU pre-integration based on flow pattern, comprising: IMU pre-integration calculates, calculates pre- score Jacobi and covariance square
Battle array calculates residual error Jacobi；
4, IMU initialization (initialization of vision inertia joint), it is one that this, which is the first step of vision and inertial data fusion,
Section loose coupling process, vision data and quick IMU data are combined, comprising: offset of gyroscope calibration (zero
Partially), scale restores and acceleration of gravity pre-estimation, accelerometer biasing calibration and scale acceleration of gravity optimize；
5, it is optimized using close coupling Optimized model, comprising: the optimization of vision inertia consecutive frame close coupling, vision inertia office
The optimization of portion's map close coupling；
Close coupling needs are added to characteristics of image in feature vector, obtain final posture information.Finally export binocular
The document of camera acquired image and IMU, time and GPS coordinate.
6, estimate that the similarity degree of pilotless automobile or robot trajectory and ambient enviroment three-dimensional map obtains real-time position
Appearance；
7, it constructs map and carries out closed loop detection using bag of words, existing map is corrected, further increases three
Tie up the precision of map.
First by IMU is estimated pose sequence and camera estimation pose sequence alignment be estimated that camera track
True scale, and the IMU pose that can predict picture frame well and last moment characteristic point are in lower frame image
Position improves signature tracking algorithmic match speed and copes with atwirl algorithm robustness, and accelerometer provides in last IMU
Gravity vector the position of estimation can be switched in the world coordinate system that practical navigation needs.Meanwhile smart phone etc. is mobile
Terminal greatly reduces the Costco Wholesale of sensor to the wilderness demand of camera.It in summary, is a kind of low-cost and high-performance
Navigation scheme.
Embodiment 2
As shown in Figure 1, the map structuring system based on GPS, IMU and binocular vision, including vehicle-mounted arithmetic center, binocular
Camera, IMU Inertial Measurement Unit, GNSS receiver based on RTK resolving etc..
Binocular camera and IMU can be integrated in whole circuit board by hardware view fusion, and whole signal transmission are all
It completes in circuit board, is directly counted between the microprocessor of control vision system and the microprocessor for controlling IMU
According to exchange, the thing of environment or target information more details is promoted to after related algorithm by visual processes, Ke Yiti
Rise the performance of whole system.The interaction and fusion that clear data can also be used pass through data between data/address bus progress sensor
Exchange.Binocular camera, IMU Inertial Measurement Unit, the data-out port based on the RTK GNSS receiver resolved pass through data
Collected data are input in vehicle-mounted arithmetic center by bus, and the storage of data is carried out in vehicle-mounted arithmetic center and is passed through
Algorithm is carried out fusion to binocular camera, IMU Inertial Measurement Unit data and to being obtained based on the GNSS receiver that RTK is resolved
GPS signal is corrected.
The vehicle-mounted arithmetic center is the core component of whole system, using ROS frame by binocular camera, IMU inertia
Measuring unit and based on RTK resolve the collected data of GNSS receiver stored and handled.It is described to be resolved based on RTK
GNSS receiver receive GPS positioning information；The binocular camera acquisition surrounding image information carries out point Yun Jianmo；It is described
IMU Inertial Measurement Unit acquisition measurement automobile triaxial attitude angle and acceleration.
The pose for initializing pilotless automobile or robot first determines pilotless automobile traveling or robot motion
Track, correct IMU Inertial Measurement Unit, the first-class data hold time consistency of binocular camera shooting；IMU inertia measurement is carried out later
The information collection and extraction of unit and binocular camera and the GNSS receiver based on RTK resolving, are sent into vehicle-mounted arithmetic center；
Secondly the IMU pre-integration based on flow pattern is carried out in vehicle-mounted arithmetic center, IMU initialization (initialization of vision inertia joint), is made
Optimized with close coupling Optimized model and is calculated；Estimate the similar of pilotless automobile or robot trajectory and ambient enviroment three-dimensional map
Degree obtains real-time pose；It constructs map and carries out closed loop detection using bag of words, existing map is corrected, into one
Step improves the precision of three-dimensional map.
Fig. 2, Fig. 3 are picture captured by binocular camera when testing in Chengdu University of Electronic Science and Technology campus, we can be with
Find out due to use binocular camera, due to know binocular camera focal length and distance and angle between the two,
A fixed signal object is taken in figure, by calculate we can the distance between more accurate judgement symbol object and observation point, mention
High detection accuracy.
The above description is only a preferred embodiment of the present invention, is not intended to restrict the invention.Obviously, those skilled in the art
Member can carry out various changes and deformation to the present invention without departing from the spirit and scope of the present invention.If in this way, of the invention
Within the scope of the claims of the present invention and its equivalent technology, then the present invention is also intended to encompass these to these modification and variation
Including change and deformation.
Claims (7)
1. a kind of map constructing method based on GPS, IMU and binocular vision, which comprises the following steps:
Step 1: the pose of initialization pilotless automobile or robot, determines pilotless automobile traveling or robot motion
Track, correct IMU Inertial Measurement Unit, the first-class data hold time consistency of binocular camera shooting；
Step 2: the information for carrying out IMU Inertial Measurement Unit and binocular camera and the GNSS receiver based on RTK resolving is adopted
Collection and extraction, to carry out visual signature tracking；
Step 3: the IMU pre-integration based on flow pattern, comprising: IMU pre-integration calculates, calculates pre- score Jacobi and covariance square
Battle array calculates residual error Jacobi；
Step 4: IMU initialization (initialization of vision inertia joint), it is one that this, which is the first step of vision and inertial data fusion,
Section loose coupling process, vision data and quick IMU data are combined, comprising: offset of gyroscope calibration (zero bias), scale are extensive
Multiple and acceleration of gravity pre-estimation, accelerometer biasing calibration and the optimization of scale acceleration of gravity；
Step 5: being optimized using close coupling Optimized model, comprising: the optimization of vision inertia consecutive frame close coupling, vision inertia
The optimization of local map close coupling；
Step 6: estimation pilotless automobile or the similarity degree of robot trajectory and ambient enviroment three-dimensional map, obtain in real time
Pose；
Step 7: constructing map and carrying out closed loop detection using bag of words, existing map is corrected, further increases three
Tie up the precision of map.
2. the map constructing method according to claim 1 based on GPS, IMU and binocular vision, which is characterized in that
In step 2) information collection and extraction:
After each binocular camera collects image, the extraction of ORB feature is carried out to image, after extracting ORB feature, will be extracted
Result storage；
The IMU data of pilotless automobile or robot interior are acquired in real time, and are sent into the IMU pre-integration based on flow pattern
In handled.
3. the map constructing method and system according to claim 1 based on GPS, IMU and binocular vision, feature exist
In in IMU pre-integration of the step 3) based on flow pattern:
During the IMU pre-integration calculates, world coordinate system, IMU coordinate system and camera coordinates system are usually with W, B, C come table
Show, Δ t is sampled at a certain time interval, measures the acceleration a of IMUBAnd angular velocity omegaB.IMU measurement model:
Wherein,The angular speed and acceleration of IMU measurement model under indicates coordinate system B,BωWB(t)∈R3It indicates
Instantaneous angular velocity of the center the IMU B relative to world coordinate system W, R under coordinate system BWBIndicate the center IMU B relative to world coordinate system
The spin matrix of W,For RWBTransposition,Wa(t)∈R3Indicate the instantaneous acceleration of the IMU under W coordinate,WG is indicated in W coordinate
It is lower gravitational vectors, b indicates zero bias, and η indicates gaussian random noise, ba、bgRespectively indicate acceleration and gravitational vectors zero bias, ηa、
ηgRespectively indicate acceleration and gravitational vectors gaussian random noise；
Rotation R, the derivative for translating p and speed v can indicate are as follows:
Wherein,WV indicates the instantaneous velocity of the IMU under W coordinate system,Indicate instantaneous angular velocity of the B relative to W, RWBIndicate IMU
Spin matrix of the center B relative to world coordinate system W,WP indicates the instantaneous translation under W coordinate system,For it
Derivative；
Rotation R, translation p and speed v under world coordinate system can be acquired by general integral formula, and Euler is used under discrete time
Integral can will above continuous time integral rewrite and simultaneous can obtain:
Δ t is certain time interval, ηgd(t)、ηad(t) indicate that acceleration and gravitational vectors gaussian random are made an uproar under discrete time
Sound, ba、bgRespectively indicate acceleration and gravitational vectors zero bias, ηad、ηgdIt respectively indicates acceleration and gravitational vectors gaussian random is made an uproar
Sound,For the angular speed and acceleration of IMU measurement model, quantity of state at this time is under world coordinate system；
If corresponding IMU quantity of state of known i moment is needed now with two adjacent key frames of i and j, when solving j under discrete time
Carve the corresponding IMU pre-integration value of key frame:
Δ t is certain time interval, Δ tijFor the time difference between i, j moment, g is gravitational vectors,It respectively indicates
The gravitational vectors and acceleration zero bias changed over time,Indicate the IMU measurement model angular speed changed over time,Indicate with
The IMU measurement model speed of time change, Rj、Ri、vj、vi、pj、piThe respectively rotation at j moment and i moment, speed and translation,
RkFor the spin matrix changed over time,Discrete acceleration and gravitational vectors gaussian random to change over time are made an uproar
Sound；
In order to avoid repetition integral is changed using the opposite of quantity of state that IMU pre-integration method calculates i, j moment, that is, will ginseng
It examines coordinate system to be transformed by world coordinate system under the i moment coordinate system of IMU, IMU quantity of state is calculated between i the and j moment
Relative quantity:
Rj、Ri、vj、vi、pj、piThe respectively rotation at j moment and i moment, speed and translation,Respectively change over time
Acceleration and gravitational vectors zero bias,Indicate the IMU measurement model angular speed changed over time,What expression changed over time
IMU measurement model speed,For the discrete acceleration and gravitational vectors gaussian random noise changed over time, Δ Rij、
Δvij、ΔpijRotation, speed and translation between i, j is poor,For Δ RiTransposition, Δ tijBetween i, j moment when
Between it is poor, Δ t be certain time interval, Δ Rik、ΔvikRotation and speed difference for i, k moment；
From the equations above as can be seen that entire pre-integration formula and biasing, noise have relationship, need to analyze respectively biasing and
Influence of the noise to system, noise obey Gauss normal distribution, bigoted obedience random walk model；
The single order item not biased in the IMU pre-integration that uses of initialization, Jacobian matrix be after the initialization is completed into
It is just calculated before row optimization, obtains the accurate pre-integration value (true value) of IMU and estimated value to obtain the residual error of IMU；
System is Gaussian distributed, and covariance matrix is calculated according to Gaussian Profile, is the matrix of a 9*9；
The calculating residual error Jacobi:
Residual error:
Parameter to be estimated:
ΔRij、Δvij、ΔpijIt is poor for the rotation at i, j moment, speed and translation,For device average value, For i
The acceleration and gravitational vectors zero bias at moment,For its average value, δ ba、δbgFor its renewal amount, φ is rotation, Δ tij
Inter frame temporal is poor, Ri、Rj、vi、vj、pi、pjIndicate spin matrix, speed and the translation at i, j moment,For its transposition；
The corresponding Jacobi of displacement residual error is as above, and the j moment, corresponding Jacobi can also so derive, and speed residual sum rotates refined
It is similar to translation than formula.So far IMU pre-integration is completed.
4. the map constructing method according to claim 1 based on GPS, IMU and binocular vision, which is characterized in that
In step 4) IMU initialization (initialization of vision inertia joint):
The rotation amount of key frame is contacted using two to estimate the biasing of gyroscope, in a constant bgOn the basis of add one
It is smallChange, calculate the renewal amount of offset of gyroscope by minimizing rotation error:
Wherein, N is the number of key frame, Δ Ri,i+1It is two continuous key frame pre-integration wheel measuring amounts, J is covariance square
Battle array,Respectively the B coordinate system at i+1 moment is relative to W coordinate system and the W coordinate system at i moment relative to B coordinate system
Spin matrix；Give zero initial value, above-mentioned function is optimized by G2O, and it is corresponding to solve minimum target functionAs gyro
The initial bias of instrument；
Pre-integration formula is substituted into after finding out zero bias can recalculate a pre-integration value, keep pre-integration numerical value more quasi-
Really；
It carries out scale to restore and acceleration of gravity pre-estimation, it is necessary first to establish pre-evaluation state vector X=[s, gW], wherein
S is scale, gWIt is the acceleration of gravity under world coordinate system；
Three key frames are used herein (indicates) that simultaneous vision and IMU pre-integration data construct an AX=B's with 1,2,3
Least square overdetermined equation at least needs four key frames, seeks least square problem using singular value decomposition, speed relatively it is slow still
Precision is high；
I is unit matrix,WPCFor the displacement of image center C under W coordinate system, RWB、RWCIt is W coordinate system relative to B coordinate system and C
The spin matrix of coordinate system, Δ tnmTime difference (n, m=1,2,3) between two frames；
Since scale restores the influence for not accounting for accelerometer biasing with the calculating process of acceleration of gravity pre-estimation, so that
Acceleration of gravity and accelerometer biasing are difficult to differentiate between, it is likely that lead to systemic disease state property problem, so it is inclined to carry out accelerometer
Set calibration and the optimization of scale acceleration of gravity；
If the acceleration of gravity direction of inertial coodinate system IgWNormalizated unit vector
θ is angle.Then the spin matrix between world coordinate system W and inertial coodinate system I is calculated:
Then acceleration of gravity is expressed as:
Due to RWIIt is rotation about the z axis, so only by X, the influence of Y-direction:
Biasing comprising acceleration:
Using three continuous key frames come release rate amount:
Wherein [](:,1:2)The first two columns of representing matrix, acceleration bias renewal amountδ θ is minute angle variable,WPCFor W seat
Mark is the displacement of lower image center C,CPBFor translation of the center the IMU B under C coordinate system, s is scale, Δ tnm、Δpnm、ΔvnmPoint
It is not time, displacement and the speed difference (n, m=1,2,3) between two frames, RWB、RWC、RWIIt is W coordinate system relative to B coordinate system, C
The spin matrix of coordinate system and I coordinate system,For g under I coordinate systemINormalizated unit vector, J is covariance matrix；It is only excellent
Change the direction of acceleration of gravity, the data that every suboptimization obtains all can calculate newest g by substituting into againWContinue to optimize, directly
It is restrained up to variable.
5. the map constructing method according to claim 1 based on GPS, IMU and binocular vision, which is characterized in that
During step 5) is optimized using close coupling Optimized model:
There are two kinds of optimal ways for the optimization of vision inertia consecutive frame close coupling: (1) there is map rejuvenation:
Global optimization state vector, the rotation comprising current time j are constructed firstTranslational velocityDisplacementAccelerate
Degree meter biasingAnd offset of gyroscopeGlobal error equation includes vision re-projection error EprojWith IMU measurement error EIMU；
Vision re-projection equation is exactly simple pinhole camera re-projection equation, and there is no what, especially its residual error, Jacobi are counted
It calculates and information matrix is no different with general vision；
IMU measuring error equation includes two parts, is P, V, R and b respectivelya,bgBiasing；Its residual error, residual error Jacobian matrix and letter
It ceases matrix and has settled accounts completion in IMU pre-integration process, be brought directly to:
eb=bj-bj (40)
eR、ep、ev、ebFor rotation, translation, speed and biased error, RWB、RWC、RWIIt is W coordinate system relative to B coordinate system, C coordinate
The spin matrix of system and I coordinate system, b are biasing；
(2) there is not map rejuvenation:
As described above, constructing global optimization state vector, the rotation including current time j+1 and last moment j first Translational velocityDisplacementAccelerometer biasingAnd offset of gyroscopeIt is whole
Error equation includes vision re-projection error Eproj, IMU measurement error EIMUWith prior uncertainty Eprior:
The error equation of vision and IMU and as before, prior uncertainty is the quantity of state at j moment；
eR、ep、ev、ebFor rotation, translation, speed and biased error,For its transposition, RWBFor W coordinate system phase
For B coordinate system spin matrix, RBWIt is B coordinate system relative to W coordinate system spin matrix, b is biasing, and ρ is coefficient,WpB、WvBFor
Translation and speed of the center the IMU B under W coordinate system；
When vision inertia local map close coupling optimizes, the key frame for needing to optimize includes that vision re-projection error and IMU are surveyed
Measure error.
6. the map constructing method according to claim 1 based on GPS, IMU and binocular vision, which is characterized in that
In the closed loop detection of step 7), constructs map and carry out closed loop detection using bag of words, existing map is corrected, into one
Step improves the precision of three-dimensional map.
7. a kind of map structuring system based on GPS, IMU and binocular vision, it is characterised in that:
The system includes vehicle-mounted arithmetic center, binocular camera, IMU Inertial Measurement Unit, the GNSS reception resolved based on RTK
Device；
The vehicle-mounted arithmetic center by ROS frame by binocular camera, IMU Inertial Measurement Unit and based on RTK resolve GNSS
The collected data of receiver are stored and are handled；
The GNSS receiver resolved based on RTK receives GPS positioning information；
The binocular camera acquisition surrounding image information carries out point Yun Jianmo；
The IMU Inertial Measurement Unit acquisition measurement automobile triaxial attitude angle and acceleration；
Binocular camera and IMU can be integrated in whole circuit board by hardware view fusion, and whole signal transmission are all in electricity
It is completed in the plate plate of road, directly carries out data between the microprocessor of control vision system and the microprocessor for controlling IMU
Exchange, is promoted the thing of environment or target information more details to after related algorithm by visual processes, can be promoted whole
The performance of a system.The interaction and fusion that clear data can also be used, the exchange of data between sensor is carried out by data/address bus.
Binocular camera, IMU Inertial Measurement Unit, the data-out port based on the RTK GNSS receiver resolved pass through data/address bus
Collected data are input in vehicle-mounted arithmetic center, the storage of data is carried out in vehicle-mounted arithmetic center and pass through algorithm
Fusion is carried out to binocular camera, IMU Inertial Measurement Unit data and to the GPS obtained based on the GNSS receiver that RTK is resolved
Signal is corrected.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910225839.8A CN109991636A (en) | 2019-03-25 | 2019-03-25 | Map constructing method and system based on GPS, IMU and binocular vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910225839.8A CN109991636A (en) | 2019-03-25 | 2019-03-25 | Map constructing method and system based on GPS, IMU and binocular vision |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109991636A true CN109991636A (en) | 2019-07-09 |
Family
ID=67131077
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910225839.8A Pending CN109991636A (en) | 2019-03-25 | 2019-03-25 | Map constructing method and system based on GPS, IMU and binocular vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109991636A (en) |
Cited By (36)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110412635A (en) * | 2019-07-22 | 2019-11-05 | 武汉大学 | A kind of environment beacon support under GNSS/SINS/ vision tight integration method |
| CN110455301A (en) * | 2019-08-01 | 2019-11-15 | 河北工业大学 | A kind of dynamic scene SLAM method based on Inertial Measurement Unit |
| CN110487267A (en) * | 2019-07-10 | 2019-11-22 | 湖南交工智能技术有限公司 | A kind of UAV Navigation System and method based on VIO&UWB pine combination |
| CN110542916A (en) * | 2019-09-18 | 2019-12-06 | 上海交通大学 | satellite and vision tight coupling positioning method, system and medium |
| CN110703754A (en) * | 2019-10-17 | 2020-01-17 | 南京航空航天大学 | Path and speed highly-coupled trajectory planning method for automatic driving vehicle |
| CN110702107A (en) * | 2019-10-22 | 2020-01-17 | 北京维盛泰科科技有限公司 | Monocular vision inertial combination positioning navigation method |
| CN110930506A (en) * | 2019-10-11 | 2020-03-27 | 深圳市道通智能航空技术有限公司 | Three-dimensional map generation method, mobile device, and computer-readable storage medium |
| CN111141295A (en) * | 2019-12-20 | 2020-05-12 | 南京航空航天大学 | Automatic map recovery method based on monocular ORB-SLAM |
| CN111337037A (en) * | 2020-05-19 | 2020-06-26 | 北京数字绿土科技有限公司 | Mobile laser radar slam drawing device and data processing method |
| CN111596665A (en) * | 2020-05-29 | 2020-08-28 | 浙江大学 | Dense height map construction method suitable for leg-foot robot planning |
| CN111623773A (en) * | 2020-07-17 | 2020-09-04 | 国汽(北京)智能网联汽车研究院有限公司 | Target positioning method and device based on fisheye vision and inertial measurement |
| CN111862150A (en) * | 2020-06-19 | 2020-10-30 | 杭州易现先进科技有限公司 | Image tracking method and device, AR device and computer device |
| CN112033422A (en) * | 2020-08-28 | 2020-12-04 | 的卢技术有限公司 | SLAM method for multi-sensor fusion |
| CN112082545A (en) * | 2020-07-29 | 2020-12-15 | 武汉威图传视科技有限公司 | Map generation method, device and system based on IMU and laser radar |
| CN112240768A (en) * | 2020-09-10 | 2021-01-19 | 西安电子科技大学 | Visual inertial navigation fusion SLAM method based on Runge-Kutta4 improved pre-integration |
| CN112256001A (en) * | 2020-09-29 | 2021-01-22 | 华南理工大学 | Visual servo control method for mobile robot under visual angle constraint |
| CN112284396A (en) * | 2020-10-29 | 2021-01-29 | 的卢技术有限公司 | Vehicle positioning method suitable for underground parking lot |
| CN112762929A (en) * | 2020-12-24 | 2021-05-07 | 华中科技大学 | Intelligent navigation method, device and equipment |
| CN112789655A (en) * | 2019-09-23 | 2021-05-11 | 北京航迹科技有限公司 | System and method for calibrating an inertial test unit and camera |
| CN112904883A (en) * | 2021-01-26 | 2021-06-04 | 德鲁动力科技(成都)有限公司 | Terrain perception method, motion control method and system for quadruped robot |
| CN113155121A (en) * | 2021-03-22 | 2021-07-23 | 珠海深圳清华大学研究院创新中心 | Vehicle positioning method and device and electronic equipment |
| WO2021147391A1 (en) * | 2020-01-21 | 2021-07-29 | 魔门塔(苏州)科技有限公司 | Map generation method and device based on fusion of vio and satellite navigation system |
| CN113190639A (en) * | 2021-05-13 | 2021-07-30 | 重庆市勘测院 | Comprehensive drawing method for residential area |
| CN113325454A (en) * | 2021-05-18 | 2021-08-31 | 武汉大学 | Combined positioning method based on ArduRover unmanned vehicle |
| CN113375665A (en) * | 2021-06-18 | 2021-09-10 | 西安电子科技大学 | Unmanned aerial vehicle pose estimation method based on multi-sensor elastic coupling |
| CN113465608A (en) * | 2021-07-22 | 2021-10-01 | 清华大学苏州汽车研究院（吴江） | Calibration method and system for roadside sensor |
| WO2021218620A1 (en) * | 2020-04-30 | 2021-11-04 | 上海商汤临港智能科技有限公司 | Map building method and apparatus, device and storage medium |
| CN113701750A (en) * | 2021-08-23 | 2021-11-26 | 长安大学 | Fusion positioning system of underground multi-sensor |
| WO2021248636A1 (en) * | 2020-06-12 | 2021-12-16 | 东莞市普灵思智能电子有限公司 | System and method for detecting and positioning autonomous driving object |
| CN114088087A (en) * | 2022-01-21 | 2022-02-25 | 深圳大学 | High-reliability high-precision navigation positioning method and system under unmanned aerial vehicle GPS-DENIED |
| CN114252099A (en) * | 2021-12-03 | 2022-03-29 | 武汉科技大学 | Intelligent vehicle multi-sensor fusion self-calibration method and system |
| CN114579679A (en) * | 2020-12-01 | 2022-06-03 | 中移(成都)信息通信科技有限公司 | Spatial positioning data fusion method, system, device and computer storage medium |
| CN115201873A (en) * | 2022-09-06 | 2022-10-18 | 中冶智诚(武汉)工程技术有限公司 | Multi-system collaborative indoor and outdoor precise positioning system architecture and operation method thereof |
| CN115307626A (en) * | 2021-05-06 | 2022-11-08 | 北京航通云科技有限公司 | Redundancy positioning method applied to small unmanned aerial vehicle |
| WO2024045632A1 (en) * | 2022-08-31 | 2024-03-07 | 华南理工大学 | Binocular vision and imu-based underwater scene three-dimensional reconstruction method, and device |
| CN113296133B (en) * | 2020-04-01 | 2024-03-15 | 易通共享技术(广州)有限公司 | Device and method for realizing position calibration based on binocular vision measurement and high-precision positioning fusion technology |
-
2019
- 2019-03-25 CN CN201910225839.8A patent/CN109991636A/en active Pending
Cited By (53)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110487267A (en) * | 2019-07-10 | 2019-11-22 | 湖南交工智能技术有限公司 | A kind of UAV Navigation System and method based on VIO&UWB pine combination |
| CN110487267B (en) * | 2019-07-10 | 2021-06-04 | 湖南交工智能技术有限公司 | Unmanned aerial vehicle navigation system and method based on VIO & UWB loose combination |
| CN110412635B (en) * | 2019-07-22 | 2023-11-24 | 武汉大学 | GNSS/SINS/visual tight combination method under environment beacon support |
| CN110412635A (en) * | 2019-07-22 | 2019-11-05 | 武汉大学 | A kind of environment beacon support under GNSS/SINS/ vision tight integration method |
| CN110455301A (en) * | 2019-08-01 | 2019-11-15 | 河北工业大学 | A kind of dynamic scene SLAM method based on Inertial Measurement Unit |
| CN110542916A (en) * | 2019-09-18 | 2019-12-06 | 上海交通大学 | satellite and vision tight coupling positioning method, system and medium |
| CN112789655A (en) * | 2019-09-23 | 2021-05-11 | 北京航迹科技有限公司 | System and method for calibrating an inertial test unit and camera |
| CN110930506B (en) * | 2019-10-11 | 2022-09-09 | 深圳市道通智能航空技术股份有限公司 | Three-dimensional map generation method, mobile device, and computer-readable storage medium |
| CN110930506A (en) * | 2019-10-11 | 2020-03-27 | 深圳市道通智能航空技术有限公司 | Three-dimensional map generation method, mobile device, and computer-readable storage medium |
| CN110703754A (en) * | 2019-10-17 | 2020-01-17 | 南京航空航天大学 | Path and speed highly-coupled trajectory planning method for automatic driving vehicle |
| CN110702107A (en) * | 2019-10-22 | 2020-01-17 | 北京维盛泰科科技有限公司 | Monocular vision inertial combination positioning navigation method |
| CN111141295A (en) * | 2019-12-20 | 2020-05-12 | 南京航空航天大学 | Automatic map recovery method based on monocular ORB-SLAM |
| WO2021147391A1 (en) * | 2020-01-21 | 2021-07-29 | 魔门塔(苏州)科技有限公司 | Map generation method and device based on fusion of vio and satellite navigation system |
| CN113296133B (en) * | 2020-04-01 | 2024-03-15 | 易通共享技术(广州)有限公司 | Device and method for realizing position calibration based on binocular vision measurement and high-precision positioning fusion technology |
| WO2021218620A1 (en) * | 2020-04-30 | 2021-11-04 | 上海商汤临港智能科技有限公司 | Map building method and apparatus, device and storage medium |
| CN111337037B (en) * | 2020-05-19 | 2020-09-29 | 北京数字绿土科技有限公司 | Mobile laser radar slam drawing device and data processing method |
| CN111337037A (en) * | 2020-05-19 | 2020-06-26 | 北京数字绿土科技有限公司 | Mobile laser radar slam drawing device and data processing method |
| CN111596665A (en) * | 2020-05-29 | 2020-08-28 | 浙江大学 | Dense height map construction method suitable for leg-foot robot planning |
| WO2021248636A1 (en) * | 2020-06-12 | 2021-12-16 | 东莞市普灵思智能电子有限公司 | System and method for detecting and positioning autonomous driving object |
| CN111862150A (en) * | 2020-06-19 | 2020-10-30 | 杭州易现先进科技有限公司 | Image tracking method and device, AR device and computer device |
| CN111623773A (en) * | 2020-07-17 | 2020-09-04 | 国汽(北京)智能网联汽车研究院有限公司 | Target positioning method and device based on fisheye vision and inertial measurement |
| CN112082545A (en) * | 2020-07-29 | 2020-12-15 | 武汉威图传视科技有限公司 | Map generation method, device and system based on IMU and laser radar |
| CN112082545B (en) * | 2020-07-29 | 2022-06-21 | 武汉威图传视科技有限公司 | Map generation method, device and system based on IMU and laser radar |
| CN112033422B (en) * | 2020-08-28 | 2023-11-21 | 的卢技术有限公司 | SLAM method for multi-sensor fusion |
| CN112033422A (en) * | 2020-08-28 | 2020-12-04 | 的卢技术有限公司 | SLAM method for multi-sensor fusion |
| CN112240768A (en) * | 2020-09-10 | 2021-01-19 | 西安电子科技大学 | Visual inertial navigation fusion SLAM method based on Runge-Kutta4 improved pre-integration |
| CN112256001A (en) * | 2020-09-29 | 2021-01-22 | 华南理工大学 | Visual servo control method for mobile robot under visual angle constraint |
| CN112284396B (en) * | 2020-10-29 | 2023-01-03 | 的卢技术有限公司 | Vehicle positioning method suitable for underground parking lot |
| CN112284396A (en) * | 2020-10-29 | 2021-01-29 | 的卢技术有限公司 | Vehicle positioning method suitable for underground parking lot |
| CN114579679A (en) * | 2020-12-01 | 2022-06-03 | 中移(成都)信息通信科技有限公司 | Spatial positioning data fusion method, system, device and computer storage medium |
| CN112762929A (en) * | 2020-12-24 | 2021-05-07 | 华中科技大学 | Intelligent navigation method, device and equipment |
| CN112762929B (en) * | 2020-12-24 | 2022-08-02 | 华中科技大学 | Intelligent navigation method, device and equipment |
| CN112904883B (en) * | 2021-01-26 | 2022-08-05 | 德鲁动力科技(成都)有限公司 | Terrain perception method, motion control method and system for quadruped robot |
| CN112904883A (en) * | 2021-01-26 | 2021-06-04 | 德鲁动力科技(成都)有限公司 | Terrain perception method, motion control method and system for quadruped robot |
| CN113155121B (en) * | 2021-03-22 | 2024-04-02 | 珠海深圳清华大学研究院创新中心 | Vehicle positioning method and device and electronic equipment |
| CN113155121A (en) * | 2021-03-22 | 2021-07-23 | 珠海深圳清华大学研究院创新中心 | Vehicle positioning method and device and electronic equipment |
| CN115307626A (en) * | 2021-05-06 | 2022-11-08 | 北京航通云科技有限公司 | Redundancy positioning method applied to small unmanned aerial vehicle |
| CN113190639B (en) * | 2021-05-13 | 2022-12-13 | 重庆市勘测院 | Comprehensive drawing method for residential area |
| CN113190639A (en) * | 2021-05-13 | 2021-07-30 | 重庆市勘测院 | Comprehensive drawing method for residential area |
| CN113325454B (en) * | 2021-05-18 | 2022-06-14 | 武汉大学 | Combined positioning method based on ArduRover unmanned vehicle |
| CN113325454A (en) * | 2021-05-18 | 2021-08-31 | 武汉大学 | Combined positioning method based on ArduRover unmanned vehicle |
| CN113375665A (en) * | 2021-06-18 | 2021-09-10 | 西安电子科技大学 | Unmanned aerial vehicle pose estimation method based on multi-sensor elastic coupling |
| CN113375665B (en) * | 2021-06-18 | 2022-12-02 | 西安电子科技大学 | Unmanned aerial vehicle pose estimation method based on multi-sensor elastic coupling |
| CN113465608A (en) * | 2021-07-22 | 2021-10-01 | 清华大学苏州汽车研究院（吴江） | Calibration method and system for roadside sensor |
| CN113465608B (en) * | 2021-07-22 | 2024-05-03 | 清华大学苏州汽车研究院（吴江） | Road side sensor calibration method and system |
| CN113701750A (en) * | 2021-08-23 | 2021-11-26 | 长安大学 | Fusion positioning system of underground multi-sensor |
| CN114252099B (en) * | 2021-12-03 | 2024-02-23 | 武汉科技大学 | Multi-sensor fusion self-calibration method and system for intelligent vehicle |
| CN114252099A (en) * | 2021-12-03 | 2022-03-29 | 武汉科技大学 | Intelligent vehicle multi-sensor fusion self-calibration method and system |
| CN114088087B (en) * | 2022-01-21 | 2022-04-15 | 深圳大学 | High-reliability high-precision navigation positioning method and system under unmanned aerial vehicle GPS-DENIED |
| CN114088087A (en) * | 2022-01-21 | 2022-02-25 | 深圳大学 | High-reliability high-precision navigation positioning method and system under unmanned aerial vehicle GPS-DENIED |
| WO2024045632A1 (en) * | 2022-08-31 | 2024-03-07 | 华南理工大学 | Binocular vision and imu-based underwater scene three-dimensional reconstruction method, and device |
| CN115201873A (en) * | 2022-09-06 | 2022-10-18 | 中冶智诚(武汉)工程技术有限公司 | Multi-system collaborative indoor and outdoor precise positioning system architecture and operation method thereof |
| CN115201873B (en) * | 2022-09-06 | 2023-07-28 | 中冶智诚(武汉)工程技术有限公司 | Multi-system cooperative indoor and outdoor precise positioning system architecture and operation method thereof |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109991636A (en) | Map constructing method and system based on GPS, IMU and binocular vision | |
| CN109631887B (en) | Inertial navigation high-precision positioning method based on binocular, acceleration and gyroscope | |
| CN108731670A (en) | Inertia/visual odometry combined navigation locating method based on measurement model optimization | |
| Panahandeh et al. | Vision-aided inertial navigation based on ground plane feature detection | |
| CN102538781B (en) | Machine vision and inertial navigation fusion-based mobile robot motion attitude estimation method | |
| CN110487267B (en) | Unmanned aerial vehicle navigation system and method based on VIO & UWB loose combination | |
| CN105953796A (en) | Stable motion tracking method and stable motion tracking device based on integration of simple camera and IMU (inertial measurement unit) of smart cellphone | |
| CN106780699A (en) | A kind of vision SLAM methods aided in based on SINS/GPS and odometer | |
| CN110146909A (en) | A kind of location data processing method | |
| CN110221328A (en) | A kind of Combinated navigation method and device | |
| CN111426320B (en) | Vehicle autonomous navigation method based on image matching/inertial navigation/milemeter | |
| CN105509739A (en) | Tightly coupled INS/UWB integrated navigation system and method adopting fixed-interval CRTS smoothing | |
| CN107728182A (en) | Flexible more base line measurement method and apparatus based on camera auxiliary | |
| CN109443348A (en) | It is a kind of based on the underground garage warehouse compartment tracking for looking around vision and inertial navigation fusion | |
| CN105953795B (en) | A kind of navigation device and method for the tour of spacecraft surface | |
| CN103776446A (en) | Pedestrian autonomous navigation calculation algorithm based on MEMS-IMU | |
| CN109931955A (en) | Strapdown inertial navigation system Initial Alignment Method based on the filtering of state correlation Lie group | |
| CN112967392A (en) | Large-scale park mapping and positioning method based on multi-sensor contact | |
| CN109764870A (en) | Carrier initial heading evaluation method based on transformation estimator modeling scheme | |
| CN115930977A (en) | Method and system for positioning characteristic degradation scene, electronic equipment and readable storage medium | |
| Klein et al. | LiDAR and INS fusion in periods of GPS outages for mobile laser scanning mapping systems | |
| Gupta et al. | Terrain‐based vehicle orientation estimation combining vision and inertial measurements | |
| CN105115507B (en) | Personal navigation system and method in a kind of double mode room based on double IMU | |
| CN113029173A (en) | Vehicle navigation method and device | |
| CN103017773B (en) | A kind of based on catalog of celestial bodies region feature and natural satellite road sign around section air navigation aid |
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 |