CN109760107A - A kind of robot localization Accuracy Assessment based on monocular vision - Google Patents
A kind of robot localization Accuracy Assessment based on monocular vision Download PDFInfo
- Publication number
- CN109760107A CN109760107A CN201910056723.6A CN201910056723A CN109760107A CN 109760107 A CN109760107 A CN 109760107A CN 201910056723 A CN201910056723 A CN 201910056723A CN 109760107 A CN109760107 A CN 109760107A
- Authority
- CN
- China
- Prior art keywords
- robot
- scaling board
- pose
- error
- coordinate system
- 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
Landscapes
- Length Measuring Devices By Optical Means (AREA)
- Manipulator (AREA)
Abstract
The invention discloses a kind of robot localization Accuracy Assessment based on monocular vision, scaling board is arranged in any position in robot motion space, in motion process, it is multiple that different location of the robot in its working environment observes same scaling board, and with such observation method multiple repairing weld, according to the data statistical characteristics of all observation informations of a motion process acquisition and observation information, the quantitative contrast of different location algorithms positioning accuracy under same working environment is realized.The present invention obtains the true pose of robot without the help of other precision instruments, and does not need a large amount of duplicate experiment tests, has saved use cost while having improved work efficiency.In addition, so that evaluation method is easier, the movement environment of robot and space are unrestricted using the scaling board of the unknown pose of environment.Finally, describing the accuracy of robot localization algorithm to be positively correlated with the expression of position error, the comparison of positioning accuracy height between algorithms of different is realized.
Description
Technical field
The present invention relates to the technical field of robot vision more particularly to a kind of robot localizations based on monocular vision
Accuracy Assessment.
Background technique
The vision positioning of mobile robot is widely used in the various aspects of Visual Navigation of Mobile Robots, the essence of positioning
Degree directly affects Mobile Robotics Navigation ability.Therefore, most important to the evaluation of localization for Mobile Robot accuracy.At present
Common evaluation method, multi-pass cross movement and obtain equipment and calibration board group of the addition with position correlation in movement environment
Carry out comparative analysis.
However, the method that arrangement movement obtains equipment and demarcates board group, application cost is high, to the transformation degree of environment
Greatly, the complexity used is increased, such evaluation method is difficult to be widely applied in practical applications.The present invention proposes one
Positioning accuracy evaluation method of the kind based on monocular vision passes through sensory feedback of the acquisition mobile robot in moving process, benefit
With the scaling board of an any position in the monocular camera observation space being mounted on robot body, different positioning are realized
The quantitative comparison of arithmetic accuracy height.
Summary of the invention
It is an object of the invention to overcome the deficiencies in the prior art, propose a kind of robot localization essence based on monocular vision
Spend evaluation method.Scaling board is arranged in this method any position in robot motion space, without demarcating between scaling board
Posture pose (position and posture).In motion process, different location of the robot in its working environment observes same scaling board
Repeatedly, and with such observation method multiple repairing weld believed according to all observation informations that a motion process acquires according to observation
The data statistical characteristics of breath realize the quantitative contrast of different location algorithms positioning accuracy under same working environment.
To achieve the above object, technical solution provided by the present invention are as follows:
A kind of robot localization Accuracy Assessment based on monocular vision, comprising the following steps:
S1: scaling board is arranged in robot motion's visual range on the way;
S2: robot a certain scaling board into space is mobile;
S3: when robot is close to the scaling board, the real-time pose information and robot of recorder people's current algorithm
With the relative pose information of the scaling board;
S4: fitting obtains error distribution curve, evaluates the height of location algorithm precision.
Further, in the step S3, the real-time pose information of robot is acquired by location algorithm.
Further, in the step S3, the relative pose information of robot and scaling board is asked by multiple view geometry method
, steps are as follows for specific calculating:
Plane of delineation coordinate system turns image pixel coordinates system:
Wherein, uO0V is image pixel coordinates system, and unit is pixel;xO1Y is plane of delineation coordinate system, and unit is millimeter;
Assuming that physical size of each pixel on u axis and v axis direction is dxAnd dy;
In above formula, dx, dy, u0, v0It is the parameter of hypothesis;
Camera coordinates system turns world coordinate system:
In above formula, (XC,YC,ZC) be camera coordinates system in picture point, (X, Y, Z) be world coordinate system in picture point;
R is 3*3 spin matrix, and t is 3*1 translation matrix, and L is expressed as 4*4 matrix;
World coordinate system and pixel coordinate system:
Wherein, f is camera focus.
Further, the step S4 fitting obtains error distribution curve, the specific steps are as follows:
PiFor the true value of the step S3 robot pose obtained, Pi' it is the pose containing error under location algorithm, calculate two
Difference between person:
Pi'=Δi·Pi;
Assuming that the error of location algorithm meets Gaussian Profile, by means of the scaling board that pose in a working environment is unknown,
By in different positions while pose P of the recorder people containing error itselfi' and its relative to scaling board pose T, to obtain
Take the position and attitude error Δ of roboti=Pi’·Ti -1·r-1, so that it is bent to obtain robot error distribution under current location algorithm
The expression of line:
Ω=Δ1·Δ2 -1=P1’·Ti -1·T2·P2 ’-1
From the above equation, we can see that obtained error expression is positively correlated with the position error of robot, therefore it can portray to be evaluated and determine
The position error distribution situation of position algorithm, to evaluate the height of location algorithm precision.
Compared with prior art, this programme principle and advantage is as follows:
1. obtaining the true pose of robot without the help of other precision instruments, and a large amount of duplicate experiments are not needed
Test, to save use cost while improve work efficiency.
2. using the scaling board of the unknown pose of environment, without obtaining posture information between multiple scaling boards in advance, so that
Evaluation method is easier, and the movement environment of robot and space are unrestricted.
3. describing the accuracy of robot localization algorithm to be positively correlated with the expression of position error, original true value is substituted
Error distribution and expression realizes the comparison of positioning accuracy height between algorithms of different.
Detailed description of the invention
Fig. 1 is a kind of flow chart of the robot localization Accuracy Assessment based on monocular vision of the present invention;
Fig. 2 is the signal that error is portrayed in a kind of robot localization Accuracy Assessment based on monocular vision of the present invention
Figure;
Fig. 3 is the schematic diagram that plane of delineation coordinate system turns image pixel coordinates system;
Fig. 4 is the schematic diagram that camera coordinates system turns world coordinate system.
Specific embodiment
The present invention is further explained in the light of specific embodiments:
Referring to shown in Fig. 1 and 2, a kind of robot localization precision evaluation side based on monocular vision described in the present embodiment
Method, comprising the following steps:
S1: scaling board is arranged in robot motion's visual range on the way;
S2: robot a certain scaling board into space is mobile;
S3: when robot is close to the scaling board, the real-time pose information and robot of recorder people's current algorithm
With the relative pose information of the scaling board;
Wherein, in step s3, the real-time pose information of robot is acquired by location algorithm.
The relative pose information of robot and scaling board is acquired by camera calibration method, and steps are as follows for specific calculating:
Plane of delineation coordinate system turns image pixel coordinates system, as shown in Figure 3:
uO0V is image pixel coordinates system, and unit is pixel;xO1Y is plane of delineation coordinate system, and unit is millimeter;
Assuming that physical size of each pixel on u axis and v axis direction is dxAnd dy;
In above formula, dx, dy, u0, v0It is the parameter of hypothesis;
Camera coordinates system turns world coordinate system, as shown in Figure 4:
Wherein, (XC,YC,ZC) be camera coordinates system in picture point, (X, Y, Z) be world coordinate system in picture point;R
For 3*3 spin matrix, t is 3*1 translation matrix, and L is expressed as 4*4 matrix;
World coordinate system and pixel coordinate system:
Wherein, f is camera focus.
S4: fitting obtains error distribution curve, evaluates the height of location algorithm precision;Specific step is as follows: PiFor step
The true value for the robot pose that rapid S3 is obtained, Pi' it is the pose containing error under location algorithm, calculate difference between the two:
Pi'=Δi·Pi;
Assuming that the error of location algorithm meets Gaussian Profile, by means of the scaling board that pose in a working environment is unknown,
By in different positions while pose P of the recorder people containing error itselfi' and its relative to scaling board pose T, to obtain
Take the position and attitude error Δ of roboti=Pi’·Ti -1·r-1, so that it is bent to obtain robot error distribution under current location algorithm
The expression of line:
Ω=Δ1·Δ2 -1=P1’·Ti -1·T2·P2’-1
From the above equation, we can see that obtained error expression is positively correlated with the position error of robot, therefore it can portray to be evaluated and determine
The position error distribution situation of position algorithm, to evaluate the height of location algorithm precision.
The present embodiment obtains the true pose of robot without the help of other precision instruments, and does not need largely to repeat
Experiment test, to save use cost while improve work efficiency.In addition, using the calibration of the unknown pose of environment
Plate, without obtaining posture information between multiple scaling boards in advance, so that evaluation method is easier, the movement environment of robot
It is unrestricted with space.Finally, describing the accuracy of robot localization algorithm, substitution to be positively correlated with the expression of position error
Original true value error distribution and expression realizes the comparison of positioning accuracy height between algorithms of different.
The examples of implementation of the above are only the preferred embodiments of the invention, and implementation model of the invention is not limited with this
It encloses, therefore all shapes according to the present invention, changes made by principle, should all be included within the scope of protection of the present invention.
Claims (4)
1. a kind of robot localization Accuracy Assessment based on monocular vision, which comprises the following steps:
S1: scaling board is arranged in robot motion's visual range on the way;
S2: robot a certain scaling board into space is mobile;
S3: when robot is close to the scaling board, the real-time pose information and robot of recorder people's current algorithm with should
The relative pose information of scaling board;
S4: fitting obtains error distribution curve, evaluates the height of location algorithm precision.
2. a kind of robot localization Accuracy Assessment based on monocular vision according to claim 1, which is characterized in that
In the step S3, the real-time pose information of robot is acquired by location algorithm.
3. a kind of robot localization Accuracy Assessment based on monocular vision according to claim 1, which is characterized in that
In the step S3, the relative pose information of robot and scaling board is acquired by camera calibration method, and steps are as follows for specific calculating:
Plane of delineation coordinate system turns image pixel coordinates system:
uO0V is image pixel coordinates system, and unit is pixel;xO1Y is plane of delineation coordinate system, and unit is millimeter;
Assuming that physical size of each pixel on u axis and v axis direction is dxAnd dy;
In above formula, dx, dy, u0, v0It is the parameter of hypothesis;
Camera coordinates system turns world coordinate system:
Wherein, (XC,YC,ZC) be camera coordinates system in picture point, (X, Y, Z) be world coordinate system in picture point;R is 3*3
Spin matrix, t are 3*1 translation matrix, and L is expressed as 4*4 matrix;
World coordinate system and pixel coordinate system:
Wherein, f is camera focus.
4. a kind of robot localization Accuracy Assessment based on monocular vision according to claim 1, which is characterized in that
The step S4 fitting obtains error distribution curve, the specific steps are as follows:
PiFor the true value of the step S3 robot pose obtained, Pi' it is the pose containing error under location algorithm, calculate the two
Between difference:
Pi'=Δi·Pi;
Assuming that the error of location algorithm meets Gaussian Profile, by means of the scaling board that pose in a working environment is unknown, pass through
In different positions while pose P of the recorder people containing error itselfi' and its relative to scaling board pose T, to obtain machine
The position and attitude error Δ of device peoplei=Pi’·Ti -1·r-1, to obtain robot error distribution curve under current location algorithm
Expression:
Ω=Δ1·Δ2 -1=P1’·Ti -1·T2·P2’-1
From the above equation, we can see that obtained error expression is positively correlated with the position error of robot, therefore evaluated positioning can be portrayed and calculated
The position error distribution situation of method, to evaluate the height of location algorithm precision.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910056723.6A CN109760107B (en) | 2019-01-22 | 2019-01-22 | Monocular vision-based robot positioning accuracy evaluation method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910056723.6A CN109760107B (en) | 2019-01-22 | 2019-01-22 | Monocular vision-based robot positioning accuracy evaluation method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109760107A true CN109760107A (en) | 2019-05-17 |
CN109760107B CN109760107B (en) | 2022-04-12 |
Family
ID=66454911
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910056723.6A Active CN109760107B (en) | 2019-01-22 | 2019-01-22 | Monocular vision-based robot positioning accuracy evaluation method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109760107B (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111012506A (en) * | 2019-12-28 | 2020-04-17 | 哈尔滨工业大学 | Robot-assisted puncture surgery end tool center calibration method based on stereoscopic vision |
CN111678521A (en) * | 2020-06-18 | 2020-09-18 | 上海大学 | Method and system for evaluating positioning accuracy of mobile robot |
CN111896032A (en) * | 2020-09-29 | 2020-11-06 | 北京清微智能科技有限公司 | Calibration system and method for monocular speckle projector position |
CN112781498A (en) * | 2021-02-09 | 2021-05-11 | 南京景曜智能科技有限公司 | Robot actuator pose repetition precision measuring method and device |
Citations (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE4415419A1 (en) * | 1994-05-02 | 1995-11-09 | Horn Wolfgang | Precision position measurement appts. for robotic container high-lift truck |
JPH11231919A (en) * | 1998-02-13 | 1999-08-27 | Hitachi Zosen Corp | Method for evaluating accuracy of robot track |
JP2000061871A (en) * | 1998-08-20 | 2000-02-29 | Hitachi Zosen Corp | Specific attitude candidate searching method for robot trajectory and precision evaluation method for robot trajectory |
CN1392950A (en) * | 2000-08-18 | 2003-01-22 | 学校法人中央大学 | Positional error evaluation method for mobile device and movement accuracy improving method based on evaluation result |
CN103389072A (en) * | 2013-07-22 | 2013-11-13 | 北京信息科技大学 | An image point positioning precision assessment method based on straight line fitting |
CN104647387A (en) * | 2013-11-25 | 2015-05-27 | 佳能株式会社 | Robot control method, system and device |
US20160193729A1 (en) * | 2015-01-06 | 2016-07-07 | Discovery Robotics | Method and system for determining precise robotic position and orientation using near-simultaneous radio frequency measurements |
CN105751245A (en) * | 2016-03-30 | 2016-07-13 | 广东工业大学 | Method and equipment for calibrating base coordinate systems of multi-robot system |
CN105945909A (en) * | 2016-05-13 | 2016-09-21 | 大族激光科技产业集团股份有限公司 | Error correction method and system for three-degree-of-freedom parallel robot |
CN106584513A (en) * | 2017-01-26 | 2017-04-26 | 武汉延锋时代检测技术股份有限公司 | Industrial robot repeated positioning accuracy testing method and industrial robot repeated positioning accuracy testing device |
CN106826922A (en) * | 2017-03-02 | 2017-06-13 | 宁波韦尔德斯凯勒智能科技有限公司 | A kind of industrial robot performance detecting system based on monocular camera machine vision |
CN108393929A (en) * | 2018-02-01 | 2018-08-14 | 大连理工大学 | A kind of mechanical arm positional precision measuring device and method |
CN108596980A (en) * | 2018-03-29 | 2018-09-28 | 中国人民解放军63920部队 | Circular target vision positioning precision assessment method, device, storage medium and processing equipment |
CN108818536A (en) * | 2018-07-12 | 2018-11-16 | 武汉库柏特科技有限公司 | A kind of online offset correction method and device of Robotic Hand-Eye Calibration |
-
2019
- 2019-01-22 CN CN201910056723.6A patent/CN109760107B/en active Active
Patent Citations (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE4415419A1 (en) * | 1994-05-02 | 1995-11-09 | Horn Wolfgang | Precision position measurement appts. for robotic container high-lift truck |
JPH11231919A (en) * | 1998-02-13 | 1999-08-27 | Hitachi Zosen Corp | Method for evaluating accuracy of robot track |
JP2000061871A (en) * | 1998-08-20 | 2000-02-29 | Hitachi Zosen Corp | Specific attitude candidate searching method for robot trajectory and precision evaluation method for robot trajectory |
CN1392950A (en) * | 2000-08-18 | 2003-01-22 | 学校法人中央大学 | Positional error evaluation method for mobile device and movement accuracy improving method based on evaluation result |
CN103389072A (en) * | 2013-07-22 | 2013-11-13 | 北京信息科技大学 | An image point positioning precision assessment method based on straight line fitting |
CN104647387A (en) * | 2013-11-25 | 2015-05-27 | 佳能株式会社 | Robot control method, system and device |
US20160193729A1 (en) * | 2015-01-06 | 2016-07-07 | Discovery Robotics | Method and system for determining precise robotic position and orientation using near-simultaneous radio frequency measurements |
CN105751245A (en) * | 2016-03-30 | 2016-07-13 | 广东工业大学 | Method and equipment for calibrating base coordinate systems of multi-robot system |
CN105945909A (en) * | 2016-05-13 | 2016-09-21 | 大族激光科技产业集团股份有限公司 | Error correction method and system for three-degree-of-freedom parallel robot |
CN106584513A (en) * | 2017-01-26 | 2017-04-26 | 武汉延锋时代检测技术股份有限公司 | Industrial robot repeated positioning accuracy testing method and industrial robot repeated positioning accuracy testing device |
CN106826922A (en) * | 2017-03-02 | 2017-06-13 | 宁波韦尔德斯凯勒智能科技有限公司 | A kind of industrial robot performance detecting system based on monocular camera machine vision |
CN108393929A (en) * | 2018-02-01 | 2018-08-14 | 大连理工大学 | A kind of mechanical arm positional precision measuring device and method |
CN108596980A (en) * | 2018-03-29 | 2018-09-28 | 中国人民解放军63920部队 | Circular target vision positioning precision assessment method, device, storage medium and processing equipment |
CN108818536A (en) * | 2018-07-12 | 2018-11-16 | 武汉库柏特科技有限公司 | A kind of online offset correction method and device of Robotic Hand-Eye Calibration |
Non-Patent Citations (3)
Title |
---|
J. WENG; N. AHUJA; T.S. HUANG: "Motion and structure from point correspondences with error estimation: planar surfaces", 《 IEEE TRANSACTIONS ON SIGNAL PROCESSING》 * |
张子淼等: "单目视觉位姿测量方法及数字仿真", 《天津大学学报》 * |
鹿玲等: "5-UPS/PRPU五自由度并联机床基准位形的确定", 《机床与液压》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111012506A (en) * | 2019-12-28 | 2020-04-17 | 哈尔滨工业大学 | Robot-assisted puncture surgery end tool center calibration method based on stereoscopic vision |
CN111678521A (en) * | 2020-06-18 | 2020-09-18 | 上海大学 | Method and system for evaluating positioning accuracy of mobile robot |
CN111678521B (en) * | 2020-06-18 | 2021-12-28 | 上海大学 | Method and system for evaluating positioning accuracy of mobile robot |
CN111896032A (en) * | 2020-09-29 | 2020-11-06 | 北京清微智能科技有限公司 | Calibration system and method for monocular speckle projector position |
CN111896032B (en) * | 2020-09-29 | 2021-09-03 | 北京清微智能科技有限公司 | Calibration system and method for monocular speckle projector position |
CN112781498A (en) * | 2021-02-09 | 2021-05-11 | 南京景曜智能科技有限公司 | Robot actuator pose repetition precision measuring method and device |
Also Published As
Publication number | Publication date |
---|---|
CN109760107B (en) | 2022-04-12 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109760107A (en) | A kind of robot localization Accuracy Assessment based on monocular vision | |
CN110377015B (en) | Robot positioning method and robot positioning device | |
JP4191080B2 (en) | Measuring device | |
KR101549103B1 (en) | Detection apparatus, Detection method and manipulator | |
CN104266608B (en) | Field calibration device for visual sensor and calibration method | |
CN105021124A (en) | Planar component three-dimensional position and normal vector calculation method based on depth map | |
JPH10253322A (en) | Method and apparatus for designating position of object in space | |
CN104766309A (en) | Plane feature point navigation and positioning method and device | |
US20220230348A1 (en) | Method and apparatus for determining a three-dimensional position and pose of a fiducial marker | |
CN112476489B (en) | Flexible mechanical arm synchronous measurement method and system based on natural characteristics | |
CN113172659B (en) | Flexible robot arm shape measuring method and system based on equivalent center point identification | |
JP2013210339A (en) | Contact state estimation device | |
Ding et al. | A robust detection method of control points for calibration and measurement with defocused images | |
CN111524174A (en) | Binocular vision three-dimensional construction method for moving target of moving platform | |
JP6626338B2 (en) | Information processing apparatus, control method for information processing apparatus, and program | |
CN111028280A (en) | # -shaped structured light camera system and method for performing scaled three-dimensional reconstruction of target | |
CN112958960A (en) | Robot hand-eye calibration device based on optical target | |
US11788831B2 (en) | Six-degree-of-freedom measurement method by machine vision based on physical decoupling | |
Zhou et al. | Visual positioning of distant wall-climbing robots using convolutional neural networks | |
CN113553920B (en) | Gaze direction feature acquisition method, gaze direction feature acquisition device, computer equipment and storage medium | |
Yamauchi et al. | Calibration of a structured light system by observing planar object from unknown viewpoints | |
KR20130075712A (en) | A laser-vision sensor and calibration method thereof | |
CN111044039B (en) | Monocular target area self-adaptive high-precision distance measurement device and method based on IMU | |
CN110415292A (en) | Movement attitude vision measurement method of ring identification and application thereof | |
CN115393299A (en) | Monocular vision-based assembly line workpiece distance measuring method and device |
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 | ||
TA01 | Transfer of patent application right | ||
TA01 | Transfer of patent application right |
Effective date of registration: 20210901 Address after: 528253 room 3, 803, floor 8, block 3, Tian'an center, No. 31, Jihua East Road, Guicheng Street, Nanhai District, Foshan City, Guangdong Province (residence declaration) Applicant after: Jiutian innovation (Guangdong) Intelligent Technology Co.,Ltd. Address before: No. 100, Waihuan West Road, University Town, Guangzhou, Guangdong 510062 Applicant before: GUANGDONG University OF TECHNOLOGY |
|
GR01 | Patent grant | ||
GR01 | Patent grant |