CN109947093A - A kind of intelligent barrier avoiding algorithm based on binocular vision - Google Patents
A kind of intelligent barrier avoiding algorithm based on binocular vision Download PDFInfo
- Publication number
- CN109947093A CN109947093A CN201910069569.6A CN201910069569A CN109947093A CN 109947093 A CN109947093 A CN 109947093A CN 201910069569 A CN201910069569 A CN 201910069569A CN 109947093 A CN109947093 A CN 109947093A
- Authority
- CN
- China
- Prior art keywords
- barrier
- robot
- point
- distance
- camera
- 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
Landscapes
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a kind of intelligent barrier avoiding algorithm based on binocular vision, include the following steps: S1, the planning in path is carried out according to the robot destination to be reached, the camera of left and right two of binocular camera is demarcated respectively, calculates separately plane equation of the ground under left and right camera coordinate system;S2 acquires robot forward image by binocular camera, and carries out denoising;S3 carries out distortion to the two images of two cameras in the left and right acquisition of binocular camera and polar curve corrects, eliminates distortion, and point-blank by match point constraint, final matching obtains a disparity map;Algorithm of the invention meets the requirement of the correct detection barrier of high-precision, and can achieve the requirement of real-time detection;The unification for meeting accuracy and real-time makes robot reach good avoidance effect.
Description
Technical field
The present invention relates to intelligent barrier avoiding fields, and in particular to a kind of intelligent barrier avoiding algorithm based on binocular vision.
Background technique
Mobile robot in actual moving process, is in dynamic uncertain environments mostly, in this environment, machine
Device people's running environment part is unknown, and some barriers in working environment are Dynamic Uncertain variations;Containing dynamic
It can only be solved by the way of online local paths planning in the environment of uncertain barrier, currently used barrier-avoiding method has
Fuzzy navigation algorithm, Artificial Potential Field algorithm, rolling window algorithm etc., still, these intelligent control algorithms do not have dynamic prediction
Function, it is poor to the avoidance effect of the dynamic barrier fast moved.Therefore forecast function is added avoidance and calculated by the prior art
In method, to improve the accuracy of dynamic obstacle avoidance, and common prediction technique includes that time series method, regression analysis, grey are pre-
Survey method etc., still, these methods mostly concentrate in the analysis to its causality regression model and time series models, thus
The model established cannot comprehensively and essence the predicted dynamic data of reflection immanent structure and complexity, to be lost letter
Breath amount.
Summary of the invention
The purpose of the present invention is to overcome the shortcomings of the existing technology with it is insufficient, a kind of intelligence based on binocular vision is provided and is kept away
Hinder algorithm, which can make robot be judged simultaneously planning path again when encountering obstacle, to achieve the effect that avoidance,
And it can satisfy the requirement of quickly judgement and real-time.
The purpose of the invention is achieved by the following technical solution:
A kind of intelligent barrier avoiding algorithm based on binocular vision, includes the following steps:
S1 carries out the planning in path according to the robot destination to be reached, the camera shooting of left and right two to binocular camera
Head is demarcated respectively;
S2 acquires robot forward image by binocular camera, and carries out denoising;
S3 carries out distortion to the two images of two cameras in the left and right acquisition of binocular camera and polar curve corrects, eliminates
Distortion, point-blank by match point constraint, final matching obtain a disparity map;
S4, calculate disparity map in three-dimensional coordinate (x, y, z) of the match point under left camera coordinate system, obtain its height with
Horizontal distance, and compared with given threshold, the distinguished number to break the barriers, which carrys out disturbance in judgement object, whether there is;Use OpenCV
The profile of cvFindCon-tours contour detecting function check object in vision library, and remembered with its boundary rectangle collimation mark;
Wherein, the distinguished number of the barrier specifically:
S4.1, it is random to extract in disparity map several white pixel point d in external rectangle framei(i=1,2,3 ...), calculates it
Three-dimensional coordinate di(xi,yi,zi), and according to horizontal distance ziSize sequence;
S4.2 it is possible that there are abnormal datas, therefore takes all whites since there are errors in extraction and calculating process
Pixel horizontal distance ziMedian z0, and given threshold φ, such as the horizontal distance z of white pixel point iiMeet | zi-z0| >
φ, then using white pixel point i as abnormity point elimination, remaining white pixel point constitutes set I;
White pixel point horizontal distance z in S4.3, set of computations IiAverage valueAs obstacle object distance in boundary rectangle frame
Horizontal distance from robot;Take in set I x in white pixel point three-dimensional coordinateiMaximum value xmaxHeight as barrier
Degree;
S4.4, setting height threshold value xdWith horizontal distance threshold value zd, when meeting xmax> xd, andWhen, then determine
Object in boundary rectangle frame is barrier;
S5, such as judgement front clear, then make robot advance according to the path of planning;Such as judge that there is obstacle in front
Object, Ze Ling robot stopping movement, and the parameter information of barrier is obtained, detailed process is as follows:
S5.1 is denoted as Q using the width of the boundary rectangle frame of barrier as the width of barrier, in boundary rectangle frame
The heart is denoted as O as barrier central point;
S5.2 obtains the three-dimensional coordinate d of barrier central point O by depth detectioni(x, y, z), wherein y indicates central point
Offset distance L relative to left camera optical centeri;A white pixel point of the leftmost side in extraneous rectangle frame is obtained in left camera shooting
Coordinate (x under head coordinate systemL,yL,zL);
S5.3 calculates the ordinate of barrier central point O according to the width of barrier, i.e. O point is relative to left camera
The offset distance L of optical centero:
S5.4, since binocular camera is mounted on the central axes of robot, two camera optical center distances are B, therefore
Offset distance d of the barrier central point O relative to robot central axes are as follows:
Wherein d>0 is indicated to the right, and d<0 indicates to the left, and d=0 indicates placed in the middle;
S6 carries out obstacle-avoiding route planning;
S6.1 sets the length of barrier as W;
S6.2, when barrier is to the left relative to robot, robot right-hand rotation angle, θ and along parallel safety distance straight trip one
Section distance X, when robot center and the line at barrier center vertical with planning path, 2 θ of left-hand rotation angle and straight trip distance X make
It returns planning path;
It is specific as follows to state shown in formula for the avoidance process of robot:
In above formula, L is that the body of robot is wide;θ and 2 θ is the rotation angle of robot;D is barrier opposed robots'
Distance;yLOffset distance for barrier left border relative to robot center;D is to consider existing error and safety problem
The increased uncompensation distance of institute.
The present invention have compared with prior art it is below the utility model has the advantages that
(1) algorithm of the invention meets the requirement of the correct detection barrier of high-precision, and can achieve real-time detection
Requirement;
(2) algorithm of the invention meets the unification of accuracy and real-time, and robot is made to have reached good avoidance effect
Fruit;
(3) algorithm of the invention has simplification, and the movement of robot obstacle-avoiding is made to have high efficiency;
(4) algorithm of the invention, which realizes, detects the barrier of different conditions in lab space, and takes corresponding
Robot dodge strategy.
Detailed description of the invention
Fig. 1 is obstacle-avoiding route planning schematic diagram of the present invention;
Fig. 2 is barrier contour detection schematic diagram of the present invention.
Specific embodiment
Present invention will now be described in further detail with reference to the embodiments and the accompanying drawings, but embodiments of the present invention are unlimited
In this.
Robot of the invention stands stop immediately, in static shape when detecting objects in front and being determined as barrier
The dimensional parameters and azimuth information of barrier are obtained when under state, and robot is controlled according to barrier dimensional parameters, azimuth information
Cut-through, specifically:
As shown in Fig. 1~2, a kind of intelligent barrier avoiding algorithm based on binocular vision includes the following steps:
Step 1 carries out the planning in path according to the robot destination to be reached, to the left and right two of binocular camera
Camera is demarcated respectively;
Step 2 acquires robot forward image, collection period T by binocular camera, and carries out denoising;
Step 3 carries out distortion to the two images of two cameras in the left and right acquisition of binocular camera and polar curve corrects,
Distortion is eliminated, point-blank by match point constraint, error hiding is reduced and greatly shortens match time, final matching obtains one
Open disparity map;
Step 4 calculates three-dimensional coordinate (x, y, z) of the match point under left camera coordinate system in disparity map, obtains its height
Degree and horizontal distance, and compared with given threshold, the distinguished number to break the barriers, which carrys out disturbance in judgement object, whether there is;It uses
The profile of cvFindCon-tours contour detecting function check object in OpenCV vision library, and with its boundary rectangle collimation mark
Note;
Wherein, the distinguished number of barrier specifically:
It extracts in disparity map in external rectangle frame several white pixel points on object at random first, calculates it and taken the photograph positioned at a left side
Three-dimensional coordinate (x, y, z) under camera coordinate system, wherein x and z respectively indicate object relative to left camera height distance with
Horizontal distance, y then indicates left and right offset distance of the object relative to left camera optical center, then according to the threshold decision square of setting
Whether the object in shape frame is barrier;
(1) several white pixel point d in external rectangle frame are extracted in disparity map at randomi(i=1,2,3 ...), calculate thirdly
Tie up coordinate di(xi,yi,zi), and according to horizontal distance zi(the z i.e. in three-dimensional coordinateiValue) size sequence;
(2) since there are errors in extraction and calculating process, it is possible that there are abnormal datas, therefore all white pictures are taken
Vegetarian refreshments horizontal distance ziMedian z0, and given threshold φ, such as the horizontal distance z of white pixel point iiMeet | zi-z0| > φ,
Then using white pixel point i as abnormity point elimination, remaining white pixel point constitutes set I;
(3) white pixel point horizontal distance z in set of computations IiAverage valueAs obstacle distance in boundary rectangle frame
The horizontal distance of robot;Take in set I x in white pixel point three-dimensional coordinateiMaximum value xmaxHeight as barrier;
(4) setting height threshold value xdWith horizontal distance threshold value zd, when meeting xmax> xd, andWhen, then determine outer
Connecing the object in rectangle frame is barrier;
Step 5, such as judgement front clear, then make robot advance according to the path of planning;Such as judge that there is barrier in front
Hinder object, Ze Ling robot stopping movement, and obtain the parameter information of barrier, for the barrier encountered during traveling,
We are other than needing to obtain depth information, it is also necessary to obtain the letter such as its width, the direction relative to our robots and offset distance
Breath, in order to which robot bypasses the obstacle, detailed process is as follows:
(1) using the width of the boundary rectangle frame of barrier as the width of barrier, Q, the center of boundary rectangle frame are denoted as
As barrier central point, it is denoted as O;
(2) the three-dimensional coordinate d of barrier central point O is obtained by depth detectioni(x, y, z), wherein y indicates central point phase
For the offset distance L of left camera optical centeri;A white pixel point of the leftmost side in extraneous rectangle frame is obtained in left camera
Coordinate (x under coordinate systemL, yL,zL);
(3) ordinate of barrier central point O is calculated according to the width of barrier, i.e. O point is relative to left camera light
The offset distance L of the hearto:
(4) since binocular camera is mounted on the central axes of robot, two camera optical center distances are B, therefore are hindered
Hinder offset distance d of the object central point O relative to robot central axes are as follows:
Wherein d>0 is indicated to the right, and d<0 indicates to the left, and d=0 indicates placed in the middle;
It can be with the specific azimuth information of acquired disturbance object by above-mentioned steps.
Step 6 carries out obstacle-avoiding route planning;
(1) length of barrier is set as W;During detection of obstacles, relatively accurately acquired disturbance is had been able to
The width dimensions of object and its azimuth information relative to robot, but it is more difficult to get the length dimension of obstacle, generally assume that it
Length is W;
(2) when barrier is to the left relative to robot, robot right-hand rotation angle, θ is simultaneously kept straight on one section along parallel safety distance
When distance X, robot center and the line at barrier center vertical with planning path, 2 θ of left-hand rotation angle and the distance X that keeps straight on make it
Return planning path;
It is specific as follows to state shown in formula for the avoidance process of robot:
In above formula, L is that the body of robot is wide;θ and 2 θ is the rotation angle of robot;D is barrier opposed robots'
Distance;yLOffset distance for barrier left border relative to robot center;D is to consider existing error and safety problem
The increased uncompensation distance of institute.
Artificial neural network has many excellent not available for conventional model method as a kind of parallel computation model
Point has good non-linear mapping capability, requires seldom, need not generally know in advance related to the Heuristics for being modeled object
It is modeled the knowledge of structure, parameter and dynamic characteristic of object etc., the input/output data of object need to be only provided, pass through
The learning functionality of network itself can reach the mapping relations of input and output.The characteristics of according to neural network, using radial direction
Basis function neural network (radial basis function network, RBFNN) establishes prediction model, and combines and roll rule
The principle drawn, solves the problems, such as the avoidance of mobile robot dynamic barrier under dynamic uncertain environments, to improve Dynamic Programming
Safety and accuracy.
Algorithm of the invention meets the requirement of the correct detection barrier of high-precision, and can achieve wanting for real-time detection
It asks;The unification for meeting accuracy and real-time makes robot reach good avoidance effect;With simplification, make machine
The movement of people's avoidance has high efficiency;It realizes and the barrier of different conditions in lab space is detected, and take corresponding
Robot dodge strategy.
Above-mentioned is the preferable embodiment of the present invention, but embodiments of the present invention are not limited by the foregoing content,
His any changes, modifications, substitutions, combinations, simplifications made without departing from the spirit and principles of the present invention, should be
The substitute mode of effect, is included within the scope of the present invention.
Claims (1)
1. a kind of intelligent barrier avoiding algorithm based on binocular vision, which is characterized in that include the following steps:
S1 carries out the planning in path according to the robot destination to be reached, to the camera of left and right two point of binocular camera
It is not demarcated;
S2 acquires robot forward image by binocular camera, and carries out denoising;
S3 carries out distortion to the two images of two cameras in the left and right acquisition of binocular camera and polar curve corrects, eliminates distortion,
Point-blank by match point constraint, final matching obtains a disparity map;
S4 calculates three-dimensional coordinate (x, y, z) of the match point under left camera coordinate system in disparity map, obtains its height and level
Distance, and compared with given threshold, the distinguished number to break the barriers, which carrys out disturbance in judgement object, whether there is;Use OpenCV vision
The profile of cvFindCon-tours contour detecting function check object in library, and remembered with its boundary rectangle collimation mark;
Wherein, the distinguished number of the barrier specifically:
S4.1, it is random to extract in disparity map several white pixel point d in external rectangle framei(i=1,2,3 ...) calculates its three-dimensional
Coordinate di(xi,yi,zi), and according to horizontal distance ziSize sequence;
S4.2 it is possible that there are abnormal datas, therefore takes all white pixels since there are errors in extraction and calculating process
Point horizontal distance ziMedian z0, and given threshold φ, such as the horizontal distance z of white pixel point iiMeet | zi-z0| > φ, then
Using white pixel point i as abnormity point elimination, remaining white pixel point constitutes set I;
White pixel point horizontal distance z in S4.3, set of computations IiAverage value z disembark as obstacle object distance in boundary rectangle frame
The horizontal distance of device people;Take in set I x in white pixel point three-dimensional coordinateiMaximum value xmaxHeight as barrier;
S4.4, setting height threshold value xdWith horizontal distance threshold value zd, when meeting xmax> xd, andWhen, then determine external
Object in rectangle frame is barrier;
S5, such as judgement front clear, then make robot advance according to the path of planning;Such as judge that there is barrier in front, then
It enables robot stopping act, and obtains the parameter information of barrier, detailed process is as follows:
S5.1 is denoted as Q using the width of the boundary rectangle frame of barrier as the width of barrier, and the center of boundary rectangle frame is made
For barrier central point, it is denoted as O;
S5.2 obtains the three-dimensional coordinate d of barrier central point O by depth detectioni(x, y, z), wherein y indicates that central point is opposite
In the offset distance L of left camera optical centeri;A white pixel point for obtaining the leftmost side in extraneous rectangle frame is sat in left camera
Coordinate (x under mark systemL,yL,zL);
S5.3 calculates the ordinate of barrier central point O according to the width of barrier, i.e. O point is relative to left camera optical center
Offset distance LO:
S5.4, since binocular camera is mounted on the central axes of robot, two camera optical centers distances are B, therefore obstacle
Offset distance d of the object central point O relative to robot central axes are as follows:
Wherein d>0 is indicated to the right, and d<0 indicates to the left, and d=0 indicates placed in the middle;
S6 carries out obstacle-avoiding route planning;
S6.1 sets the length of barrier as W;
S6.2, when barrier is to the left relative to robot, robot right-hand rotation angle, θ and along parallel safety distance straight trip one section away from
From X, when robot center and the line at barrier center vertical with planning path, 2 θ of left-hand rotation angle and the distance X that keeps straight on make its time
Return planning path;
It is specific as follows to state shown in formula for the avoidance process of robot:
In above formula, L is that the body of robot is wide;θ and 2 θ is the rotation angle of robot;D is the distance of barrier opposed robots;
yLOffset distance for barrier left border relative to robot center;D error and safety problem existing for consideration increases
The uncompensation distance added.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910069569.6A CN109947093A (en) | 2019-01-24 | 2019-01-24 | A kind of intelligent barrier avoiding algorithm based on binocular vision |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910069569.6A CN109947093A (en) | 2019-01-24 | 2019-01-24 | A kind of intelligent barrier avoiding algorithm based on binocular vision |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109947093A true CN109947093A (en) | 2019-06-28 |
Family
ID=67007252
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910069569.6A Pending CN109947093A (en) | 2019-01-24 | 2019-01-24 | A kind of intelligent barrier avoiding algorithm based on binocular vision |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109947093A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110378915A (en) * | 2019-07-24 | 2019-10-25 | 西南石油大学 | A kind of climbing robot obstacle detection method based on binocular vision |
CN110825094A (en) * | 2019-12-04 | 2020-02-21 | 安徽理工大学 | Robot patrols line navigation based on machine vision |
CN111399505A (en) * | 2020-03-13 | 2020-07-10 | 浙江工业大学 | Mobile robot obstacle avoidance method based on neural network |
CN111982058A (en) * | 2020-08-04 | 2020-11-24 | 北京中科慧眼科技有限公司 | Distance measurement method, system and equipment based on binocular camera and readable storage medium |
CN112051853A (en) * | 2020-09-18 | 2020-12-08 | 哈尔滨理工大学 | Intelligent obstacle avoidance system and method based on machine vision |
WO2021238000A1 (en) * | 2020-05-29 | 2021-12-02 | 苏州科瓴精密机械科技有限公司 | Boundary-following working method and system of robot, robot, and readable storage medium |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102902271A (en) * | 2012-10-23 | 2013-01-30 | 上海大学 | Binocular vision-based robot target identifying and gripping system and method |
CN104484648A (en) * | 2014-11-27 | 2015-04-01 | 浙江工业大学 | Variable-viewing angle obstacle detection method for robot based on outline recognition |
US20160307052A1 (en) * | 2015-04-16 | 2016-10-20 | Electronics And Telecommunications Research Institute | Device and method for recognizing obstacle and parking slot to support unmanned autonomous parking function |
CN106681353A (en) * | 2016-11-29 | 2017-05-17 | 南京航空航天大学 | Unmanned aerial vehicle (UAV) obstacle avoidance method and system based on binocular vision and optical flow fusion |
CN108345823A (en) * | 2017-01-23 | 2018-07-31 | 郑州宇通客车股份有限公司 | A kind of barrier tracking and device based on Kalman filtering |
-
2019
- 2019-01-24 CN CN201910069569.6A patent/CN109947093A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102902271A (en) * | 2012-10-23 | 2013-01-30 | 上海大学 | Binocular vision-based robot target identifying and gripping system and method |
CN104484648A (en) * | 2014-11-27 | 2015-04-01 | 浙江工业大学 | Variable-viewing angle obstacle detection method for robot based on outline recognition |
US20160307052A1 (en) * | 2015-04-16 | 2016-10-20 | Electronics And Telecommunications Research Institute | Device and method for recognizing obstacle and parking slot to support unmanned autonomous parking function |
CN106681353A (en) * | 2016-11-29 | 2017-05-17 | 南京航空航天大学 | Unmanned aerial vehicle (UAV) obstacle avoidance method and system based on binocular vision and optical flow fusion |
CN108345823A (en) * | 2017-01-23 | 2018-07-31 | 郑州宇通客车股份有限公司 | A kind of barrier tracking and device based on Kalman filtering |
Non-Patent Citations (1)
Title |
---|
王铮 等: "基于双目视觉的AGV障碍物检测与避障", 《计算机集成制造系统》 * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110378915A (en) * | 2019-07-24 | 2019-10-25 | 西南石油大学 | A kind of climbing robot obstacle detection method based on binocular vision |
CN110825094A (en) * | 2019-12-04 | 2020-02-21 | 安徽理工大学 | Robot patrols line navigation based on machine vision |
CN111399505A (en) * | 2020-03-13 | 2020-07-10 | 浙江工业大学 | Mobile robot obstacle avoidance method based on neural network |
CN111399505B (en) * | 2020-03-13 | 2023-06-30 | 浙江工业大学 | Mobile robot obstacle avoidance method based on neural network |
WO2021238000A1 (en) * | 2020-05-29 | 2021-12-02 | 苏州科瓴精密机械科技有限公司 | Boundary-following working method and system of robot, robot, and readable storage medium |
CN111982058A (en) * | 2020-08-04 | 2020-11-24 | 北京中科慧眼科技有限公司 | Distance measurement method, system and equipment based on binocular camera and readable storage medium |
CN112051853A (en) * | 2020-09-18 | 2020-12-08 | 哈尔滨理工大学 | Intelligent obstacle avoidance system and method based on machine vision |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109947093A (en) | A kind of intelligent barrier avoiding algorithm based on binocular vision | |
CN110349250B (en) | RGBD camera-based three-dimensional reconstruction method for indoor dynamic scene | |
CN110569704B (en) | Multi-strategy self-adaptive lane line detection method based on stereoscopic vision | |
CN106681353B (en) | The unmanned plane barrier-avoiding method and system merged based on binocular vision with light stream | |
CN104700414B (en) | A kind of road ahead pedestrian's fast ranging method based on vehicle-mounted binocular camera | |
CN107833236A (en) | Semantic vision positioning system and method are combined under a kind of dynamic environment | |
Brand et al. | Stereo-vision based obstacle mapping for indoor/outdoor SLAM | |
EP2757524B1 (en) | Depth sensing method and system for autonomous vehicles | |
CN109472831A (en) | Obstacle recognition range-measurement system and method towards road roller work progress | |
CN103925920A (en) | Image perspective-based micro unmanned aerial vehicle indoor autonomous navigation method | |
CN111521195B (en) | Intelligent robot | |
CN113903011B (en) | Semantic map construction and positioning method suitable for indoor parking lot | |
Moghadam et al. | Road direction detection based on vanishing-point tracking | |
CN111998862B (en) | BNN-based dense binocular SLAM method | |
CN109471096A (en) | Multi-Sensor Target matching process, device and automobile | |
CN112308921B (en) | Combined optimization dynamic SLAM method based on semantics and geometry | |
Chen | Monocular vision-based obstacle detection and avoidance for a multicopter | |
CN112446882A (en) | Robust visual SLAM method based on deep learning in dynamic scene | |
CN111258311A (en) | Obstacle avoidance method of underground mobile robot based on intelligent vision | |
CN109764869A (en) | A kind of positioning of autonomous crusing robot and the three-dimensional map construction method of binocular camera and inertial navigation fusion | |
CN108492324A (en) | Aircraft method for tracing based on fully-connected network and Kalman filter | |
CN110147748A (en) | A kind of mobile robot obstacle recognition method based on road-edge detection | |
WO2024114119A1 (en) | Sensor fusion method based on binocular camera guidance | |
Bozorgi et al. | 2D laser and 3D camera data integration and filtering for human trajectory tracking | |
WO2021063756A1 (en) | Improved trajectory estimation based on ground truth |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190628 |
|
RJ01 | Rejection of invention patent application after publication |