CN111308993B - Human body target following method based on monocular vision - Google Patents
Human body target following method based on monocular vision Download PDFInfo
- Publication number
- CN111308993B CN111308993B CN202010089860.2A CN202010089860A CN111308993B CN 111308993 B CN111308993 B CN 111308993B CN 202010089860 A CN202010089860 A CN 202010089860A CN 111308993 B CN111308993 B CN 111308993B
- Authority
- CN
- China
- Prior art keywords
- target
- frame
- tracking
- human body
- robot
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 24
- 238000001514 detection method Methods 0.000 claims abstract description 58
- 241000195940 Bryophyta Species 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000004069 differentiation Effects 0.000 claims description 3
- 238000006073 displacement reaction Methods 0.000 claims description 3
- 239000000284 extract Substances 0.000 claims description 3
- 230000010354 integration Effects 0.000 claims description 3
- 230000004044 response Effects 0.000 abstract description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/0094—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots involving pointing a payload, e.g. camera, weapon, sensor, towards a fixed or moving target
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B11/00—Automatic controllers
- G05B11/01—Automatic controllers electric
- G05B11/36—Automatic controllers electric with provision for obtaining particular characteristics, e.g. proportional, integral, differential
- G05B11/42—Automatic controllers electric with provision for obtaining particular characteristics, e.g. proportional, integral, differential for obtaining a characteristic which is both proportional and time-dependent, e.g. P. I., P. I. D.
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/12—Target-seeking control
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/223—Analysis of motion using block-matching
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30196—Human being; Person
Landscapes
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Aviation & Aerospace Engineering (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Human Computer Interaction (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Manipulator (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a human body target following method based on monocular vision, which comprises the following steps: starting a following system of the monocular vision robot under the voice control, and starting a following function of the robot; detecting a target right in front, and detecting a plurality of human body targets right in front of the robot by using a human body recognition algorithm; selecting a target to be followed, and selecting a human body target which best meets the requirements from the obtained human body detection frames as the target to be followed; setting a tracker state, and acquiring tracking information of a target to be followed in a mode of combining a main tracker and an auxiliary tracker; motion following control, namely setting a corresponding following strategy according to the relation between the target information and the following information; and updating the following target information, and using the tracking information at the current moment as the target information at the next moment to realize continuous tracking. The method disclosed by the invention adopts a monocular camera to perform an algorithm of human body target following, and has the characteristics of strong real-time performance, high detection rate and quicker response.
Description
Technical Field
The invention relates to the field of computer vision and robot motion control, in particular to a human body target following method based on monocular vision.
Background
At present, the human body target following technology of the robot is mostly developed based on a 3D depth sensor or based on a combination of monocular and distance sensors. Compared with a monocular sensor, the 3D sensor has the advantages of high cost which is about 10 times higher, large volume and incapability of being applied to a miniature robot. Moreover, the technical difficulty of data fusion of the monocular camera and the distance sensor is high, the constructed system is redundant, and the use of a small robot cannot be met.
When a robot monocular vision human body following algorithm is developed, the following problems need to be solved:
1. determining a proper target detection algorithm to guarantee the real-time performance of target detection and the difficult problem of retrieving the target again after the target is shielded;
2. developing a motion control following algorithm, and determining a motion control method and control parameters after the target moves;
3. and realizing the linkage of the voice recognition and the following algorithm.
Disclosure of Invention
In order to solve the technical problems, the invention provides a human body target following method based on monocular vision, an algorithm for following the human body target can be carried out by adopting a monocular camera, and the method has the characteristics of strong real-time performance, high detection rate and quicker response.
In order to achieve the purpose, the technical scheme of the invention is as follows:
a human body target following method based on monocular vision comprises the following steps:
step one, starting a following system of the monocular vision robot through voice control, and starting a following function of the robot;
detecting a target right in front, and detecting a plurality of human body targets right in front of the robot by using a human body recognition algorithm;
selecting a target to be followed, and selecting a human body target which best meets the requirements from the obtained human body detection frames as the target to be followed;
setting the state of a tracker, and acquiring tracking information of the target to be followed by using a mode of combining a main tracker and an auxiliary tracker;
step five, motion following control, namely setting a corresponding following strategy according to the relation between the target information and the following information;
and step six, updating the following target information, and using the tracking information at the current moment as the target information at the next moment to realize continuous tracking.
In the above scheme, the specific method of the second step is as follows:
(1) obtaining a human target detection frame using a human recognition algorithm
The robot issues voice, shoots a front part by using a monocular camera, detects each shot frame, extracts posture information of each human body to construct a rectangular human body target detection frame, the coordinates of the upper left corner of the detection frame are (x, y), and the width and the height of the rectangular frame are w and h respectively;
(2) detection frame for rejecting errors
The width of the image shot by the monocular camera is W, the height of the image is H, the wrong detection frame is removed according to the relative size of the detection frame in the shot image, and the human body target detection frame is preliminarily screened.
Further, only the width w of the rectangular frame is kept larger thanAnd is less thanThe human target detection frame.
In the above scheme, the specific method of the third step is as follows:
selecting a target detection frame with a larger area and a center coordinate close to the center coordinate of the image from a plurality of rectangular detection frames, and calculating a score S for each rectangular detection frame:
wherein σ2In order to self-define the coefficient,
selecting a rectangular detection frame with the highest score S as a target to be followed;
detecting whether the target to be followed is right in front of the robot whenAnd then, issuing a voice prompt to remind the target to move to the central position, and recording the rectangular frame information of the target to be tracked.
In the above scheme, the specific method of the fourth step is as follows:
the human body tracking function is realized by combining a main tracker MEDIANFLOW with a stable anti-loss auxiliary tracker MOSSE;
(1) tracking state detection
Detecting a tracking target in the updated image frame, setting the tracking state as true when the tracking target is found in the updated image frame, and drawing a tracking frame;
(2) drawing tracking frame
Drawing a tracking frame r on a new image frame F, wherein a main tracker corresponds to a main tracking frame r _ m, and an auxiliary tracker corresponds to an auxiliary tracking frame r _ a; drawing only one tracking frame on the updated image frame, using a main tracking frame r _ m when the main tracker is available, and using an auxiliary tracking frame r _ a when the main tracker is unavailable;
(3) detecting incorrect tracking information and resetting the tracker
Comparing the previous time information of the target to be tracked with the current information calculated by the tracker, and judging the tracking state; the information at the previous moment is represented on the image as a target frame, and the information tracked at the current moment is represented as a tracking frame; the coordinates of the top left vertex of the target frame in the image are (x, y), the width is w, the height is h, and the coordinates of the center point of the target frame areDrawing the coordinate of the upper left vertex of the tracking frame of the moving target in the image as (x) by using the trackerr,yr) And has a width of wrHeight of hrThe coordinate of the central point of the tracking frame is
Calculating the distance D between the center point of the target frame and the center point of the tracking frame:
In the above scheme, the concrete method of the fifth step is as follows:
(1) proximity detection
When tracking the height h of the framerWhen the height difference with the rectangular frame detected by the human body detection algorithm at the beginning is smaller than a set threshold value, the robot state is set to be closer to the following target, and the robot does not move forwards;
(2) normal tracking
The rotational speed rw of the rotational movement of the robot is calculated as follows:
where W is the width of the image, σ1To define the parameters, rwmaxIs a preset maximum rotation angular velocity;
the linear velocity lv calculation formula of the robot is as follows:
where W is the width of the image, σ2To define the parameters, lvmaxIs a preset maximum linear velocity;
when tracking the height h of the framerWhen the height h between the robot and the target frame is smaller than a micro movement threshold value, only calculating the rotation angular velocity of the robot, and calculating the parameters of a PID controller;
when tracking the height h of the framerAnd when the height h of the target frame is larger than the micro movement threshold, simultaneously calculating the rotation angular velocity and the displacement linear velocity of the robot, and calculating the parameters of the PID controller.
(3) Lost finding
When tracing the abscissa x of the top left vertex of the framer<k1OrJudging the current state as lost, adjusting the motion state of the robot to be a pure rotation state at the moment, calculating the parameters of a PID (proportion integration differentiation) controller, and if the parameters are more than a direction control threshold thrpThe angular velocity is set to rotate counterclockwise, if less than a direction control threshold such as thrnThen the angular velocity is set to rotate clockwise, at which time the robot rotates to retrieve the following target.
Through the technical scheme, the human body target following method based on monocular vision has the following beneficial effects:
1. setting a following function of the specific voice awakening robot, and simplifying man-machine interaction;
2. selecting and combining proper human target detection algorithms, and simultaneously ensuring the accuracy and real-time performance of human target detection;
3. the algorithm of the tracking system is optimized, and the human body target can be followed by only one camera, so that the coupling degree of the system is simplified, and the fault tolerance of the system is increased;
4. and a tracking strategy is optimized, so that the target can be tracked in time when the human target moves, and the target cannot be lost when the vehicle turns.
5. The motion control algorithm is optimized, the optimal PID parameters are calculated, and the rapidness, the accuracy and the stability of motion control are improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below.
FIG. 1 is a flowchart of a monocular vision-based human target following method according to an embodiment of the present invention;
fig. 2 is a schematic diagram of a target detection block.
Detailed Description
The technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.
The invention provides a human body target following method based on monocular vision, as shown in figure 1, the specific embodiment is as follows:
a human body target following method based on monocular vision comprises the following steps:
step one, starting a following system of the monocular vision robot through voice control, and starting a following function of the robot;
the user carries out voice conversation with the robot, and when the robot recognizes the voice which is set by the system and is related to the following instruction, the following system is started.
Detecting a target right in front, and detecting a plurality of human body targets right in front of the robot by using a human body recognition algorithm;
(1) obtaining a human detection box using a human recognition algorithm
The robot issues voice of 'please stand in front of me' and shoots the front right by using a monocular camera, detects each frame shot, extracts posture information of each human body in the frame to construct a rectangular human body target detection frame, as shown in fig. 2, the coordinates of the upper left corner of the human body target detection frame are (x, y), and the width and the height of the detection frame are w and h respectively;
(2) detection frame for rejecting errors
The width of an image shot by the monocular camera is W, the height of the image is H, wrong detection frames are removed according to the relative size of the human body detection frames in the shot image, and the human body target detection frames are preliminarily screened. Only the width w of the rectangular frame is kept larger thanAnd is less thanThe detection frame of (1).
Selecting a target to be followed, and selecting a human body target which best meets the requirements from the obtained human body detection frames as the target to be followed;
selecting a target detection frame with a larger area and a center coordinate close to the center coordinate of the image from a plurality of rectangular detection frames, and calculating a score S for each rectangular detection frame:
wherein σ2In order to self-define the coefficient,
selecting a rectangular detection frame with the highest score S as a target to be followed;
detecting whether the target to be followed is right in front of the robot whenAnd then, issuing a voice prompt to remind the target to move to the central position, and recording the rectangular frame information of the target to be tracked.
Setting the state of a tracker, and acquiring tracking information of the target to be followed by using a mode of combining a main tracker and an auxiliary tracker;
the human body tracking function is realized by combining a main tracker MEDIANFLOW with a stable anti-loss auxiliary tracker MOSSE;
(1) tracking state detection
Detecting a tracking target in the updated image frame, setting the tracking state as true when the tracking target is found in the updated image frame, and drawing a tracking frame;
(2) drawing tracking frame
Drawing a tracking frame r on a new image frame F, wherein a main tracker corresponds to a main tracking frame r _ m, and an auxiliary tracker corresponds to an auxiliary tracking frame r _ a; drawing only one tracking frame on the updated image frame, using a main tracking frame r _ m when the main tracker is available, and using an auxiliary tracking frame r _ a when the main tracker is unavailable;
(3) detecting incorrect tracking information and resetting the tracker
Comparing the previous time information of the target to be tracked with the current information calculated by the tracker, and judging the tracking state; the information of the previous moment is represented on the image as a targetThe frame, the information tracked at the present moment is represented as a tracking frame; the coordinates of the top left vertex of the target frame in the image are (x, y), the width is w, the height is h, and the coordinates of the center point of the target frame areDrawing the coordinate of the upper left vertex of the tracking frame of the moving target in the image as (x) by using the trackerr,yr) And has a width of wrHeight of hrThe coordinate of the central point of the tracking frame is
Calculating the distance D between the center point of the target frame and the center point of the tracking frame:
Step five, motion following control, namely setting a corresponding following strategy according to the relation between the target information and the following information;
(1) proximity detection
When tracking the height h of the framerWhen the height difference with the rectangular frame detected by the human body detection algorithm at the beginning is smaller than a set threshold value, the robot state is set to be closer to the following target, and the robot does not move forwards;
(2) normal tracking
The rotational speed rw of the rotational movement of the robot is calculated as follows:
where W is the width of the image, σ1To define the parameters, rwmaxAt a preset maximum rotationAn angular velocity;
the linear velocity lv calculation formula of the robot is as follows:
where W is the width of the image, σ2To define the parameters, lvmaxIs a preset maximum linear velocity;
when tracking the height h of the framerWhen the height h of the target frame is less than a micro movement threshold value 10, only calculating the rotation angular velocity of the robot, and calculating the parameters of a PID controller;
when tracking the height h of the framerAnd when the height h of the target frame is greater than the micro movement threshold value 10, simultaneously calculating the rotation angular velocity and the displacement linear velocity of the robot, and calculating the parameters of the PID controller.
(3) Lost finding
When tracing the abscissa x of the top left vertex of the framer<k1OrJudging the current state as lost, adjusting the motion state of the robot to be a pure rotation state at the moment, calculating the parameters of a PID (proportion integration differentiation) controller, and if the parameters are more than a direction control threshold thrpThen the angular velocity is set to rotate counterclockwise, e.g., a rad/s, if less than the directional control threshold, e.g., thrnThen the angular velocity is set to rotate clockwise-a rad/s, at which time the robot rotates to retrieve the following target.
And step six, updating the following target information, and using the tracking information at the current moment as the target information at the next moment to realize continuous tracking.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (3)
1. A human body target following method based on monocular vision is characterized by comprising the following steps:
step one, starting a following system of the monocular vision robot through voice control, and starting a following function of the robot;
detecting a target right in front, and detecting a plurality of human body targets right in front of the robot by using a human body recognition algorithm;
selecting a target to be followed, and selecting a human body target which best meets the requirements from the obtained human body detection frames as the target to be followed;
setting the state of a tracker, and acquiring tracking information of the target to be followed by using a mode of combining a main tracker and an auxiliary tracker;
step five, motion following control, namely setting a corresponding following strategy according to the relation between the target information and the following information;
step six, updating the following target information, and using the tracking information at the current moment as the target information at the next moment to realize continuous tracking;
the specific method of the second step is as follows:
(1) obtaining a human target detection frame using a human recognition algorithm
The robot issues voice, shoots a front part by using a monocular camera, detects each shot frame, extracts posture information of each human body to construct a rectangular human body target detection frame, the coordinates of the upper left corner of the detection frame are (x, y), and the width and the height of the detection frame are w and h respectively;
(2) detection frame for rejecting errors
The width of an image shot by the monocular camera is W, the height of the image is H, wrong detection frames are removed according to the relative size of the detection frames in the shot image, and human body target detection frames are preliminarily screened;
the concrete method of the third step is as follows:
selecting a target detection frame with a larger area and a center coordinate close to the center coordinate of the image from a plurality of rectangular detection frames, and calculating a score S for each rectangular detection frame:
wherein σ2In order to self-define the coefficient,
selecting a rectangular detection frame with the highest score S as a target to be followed;
detecting whether the target to be followed is right in front of the robot whenWhen the target is moved to the central position, issuing voice to remind the target to move to the central position, and recording rectangular frame information of the target to be tracked;
the concrete method of the step five is as follows:
(1) proximity detection
When tracking the height h of the framerWhen the height difference with the rectangular frame detected by the human body detection algorithm at the beginning is smaller than a set threshold value, the robot state is set to be closer to the following target, and the robot does not move forwards;
(2) normal tracking
The rotational speed rw of the rotational movement of the robot is calculated as follows:
where W is the width of the image, σ1To define the parameters, rwmaxIs a preset maximum rotation angular velocity;
the linear velocity lv calculation formula of the robot is as follows:
where W is the width of the image, σ2To define the parameters, lvmaxIs a preset maximum linear velocity;
when tracking the height h of the framerWhen the height h between the robot and the target frame is smaller than a micro movement threshold value, only calculating the rotation angular velocity of the robot, and calculating the parameters of a PID controller;
when tracking the height h of the framerWhen the height h of the target frame is larger than the micro moving threshold, simultaneously calculating the rotation angular velocity and the displacement linear velocity of the robot, and calculating the parameters of a PID controller;
(3) lost finding
When tracing the abscissa x of the top left vertex of the framer<k1OrJudging the current state as lost, adjusting the motion state of the robot to be a pure rotation state at the moment, calculating the parameters of a PID (proportion integration differentiation) controller, and if the parameters are more than a direction control threshold thrpThe angular velocity is set to rotate counterclockwise, if less than a direction control threshold such as thrnThen the angular velocity is set to rotate clockwise, at which time the robot rotates to retrieve the following target.
3. The human body target following method based on the monocular vision as claimed in claim 1, wherein the specific method of the fourth step is as follows:
the human body tracking function is realized by combining a main tracker MEDIANFLOW with a stable anti-loss auxiliary tracker MOSSE;
(1) tracking state detection
Detecting a tracking target in the updated image frame, setting the tracking state as true when the tracking target is found in the updated image frame, and drawing a tracking frame;
(2) drawing tracking frame
Drawing a tracking frame r on a new image frame F, wherein a main tracker corresponds to a main tracking frame r _ m, and an auxiliary tracker corresponds to an auxiliary tracking frame r _ a; drawing only one tracking frame on the updated image frame, using a main tracking frame r _ m when the main tracker is available, and using an auxiliary tracking frame r _ a when the main tracker is unavailable;
(3) detecting incorrect tracking information and resetting the tracker
Comparing the previous time information of the target to be tracked with the current information calculated by the tracker, and judging the tracking state; the information at the previous moment is represented on the image as a target frame, and the information tracked at the current moment is represented as a tracking frame; the coordinates of the top left vertex of the target frame in the image are (x, y), the width is w, the height is h, and the coordinates of the center point of the target frame areDrawing the coordinate of the upper left vertex of the tracking frame of the moving target in the image as (x) by using the trackerr,yr) And has a width of wrHeight of hrThe coordinate of the central point of the tracking frame is
Calculating the distance D between the center point of the target frame and the center point of the tracking frame:
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010089860.2A CN111308993B (en) | 2020-02-13 | 2020-02-13 | Human body target following method based on monocular vision |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010089860.2A CN111308993B (en) | 2020-02-13 | 2020-02-13 | Human body target following method based on monocular vision |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111308993A CN111308993A (en) | 2020-06-19 |
CN111308993B true CN111308993B (en) | 2022-04-01 |
Family
ID=71148974
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010089860.2A Active CN111308993B (en) | 2020-02-13 | 2020-02-13 | Human body target following method based on monocular vision |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111308993B (en) |
Families Citing this family (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111862154B (en) * | 2020-07-13 | 2024-03-01 | 中移(杭州)信息技术有限公司 | Robot vision tracking method and device, robot and storage medium |
CN111881827B (en) * | 2020-07-28 | 2022-04-26 | 浙江商汤科技开发有限公司 | Target detection method and device, electronic equipment and storage medium |
CN112207821B (en) * | 2020-09-21 | 2021-10-01 | 大连遨游智能科技有限公司 | Target searching method of visual robot and robot |
CN112132864B (en) * | 2020-09-21 | 2024-04-09 | 大连遨游智能科技有限公司 | Vision-based robot following method and following robot |
CN112880557B (en) * | 2021-01-08 | 2022-12-09 | 武汉中观自动化科技有限公司 | Multi-mode tracker system |
CN113221754A (en) * | 2021-05-14 | 2021-08-06 | 深圳前海百递网络有限公司 | Express waybill image detection method and device, computer equipment and storage medium |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2010073432A1 (en) * | 2008-12-24 | 2010-07-01 | 株式会社ソニー・コンピュータエンタテインメント | Image processing device and image processing method |
CN103942672A (en) * | 2014-05-19 | 2014-07-23 | 北京玛施德利科技有限公司 | Tracking device for logistics tracking system as well as application method of tracking device and logistics tracking system |
WO2017147792A1 (en) * | 2016-03-01 | 2017-09-08 | SZ DJI Technology Co., Ltd. | Methods and systems for target tracking |
CN108646761A (en) * | 2018-07-12 | 2018-10-12 | 郑州大学 | Robot indoor environment exploration, avoidance and method for tracking target based on ROS |
CN108875683A (en) * | 2018-06-30 | 2018-11-23 | 北京宙心科技有限公司 | Robot vision tracking method and system |
CN109741369A (en) * | 2019-01-03 | 2019-05-10 | 北京邮电大学 | A kind of method and system for robotic tracking target pedestrian |
-
2020
- 2020-02-13 CN CN202010089860.2A patent/CN111308993B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2010073432A1 (en) * | 2008-12-24 | 2010-07-01 | 株式会社ソニー・コンピュータエンタテインメント | Image processing device and image processing method |
CN103942672A (en) * | 2014-05-19 | 2014-07-23 | 北京玛施德利科技有限公司 | Tracking device for logistics tracking system as well as application method of tracking device and logistics tracking system |
WO2017147792A1 (en) * | 2016-03-01 | 2017-09-08 | SZ DJI Technology Co., Ltd. | Methods and systems for target tracking |
CN108875683A (en) * | 2018-06-30 | 2018-11-23 | 北京宙心科技有限公司 | Robot vision tracking method and system |
CN108646761A (en) * | 2018-07-12 | 2018-10-12 | 郑州大学 | Robot indoor environment exploration, avoidance and method for tracking target based on ROS |
CN109741369A (en) * | 2019-01-03 | 2019-05-10 | 北京邮电大学 | A kind of method and system for robotic tracking target pedestrian |
Non-Patent Citations (1)
Title |
---|
基于视觉目标跟踪的侦察机器人导航方法;包加桐等;《东南大学学报(自然科学版)》;20120520(第03期);全文 * |
Also Published As
Publication number | Publication date |
---|---|
CN111308993A (en) | 2020-06-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111308993B (en) | Human body target following method based on monocular vision | |
CN109631896B (en) | Parking lot autonomous parking positioning method based on vehicle vision and motion information | |
EP2460629B1 (en) | Control method for localization and navigation of mobile robot and mobile robot using same | |
JP4931218B2 (en) | Imaging apparatus, object detection method, and attitude parameter calculation method | |
WO2018159168A1 (en) | System and method for virtually-augmented visual simultaneous localization and mapping | |
CN106767833B (en) | A kind of robot localization method merging RGBD depth transducer and encoder | |
CN112669354A (en) | Multi-camera motion state estimation method based on vehicle incomplete constraint | |
Smith et al. | Eye-in-hand robotic tasks in uncalibrated environments | |
CN116030099A (en) | PTZ camera-based multi-target tracking method and device | |
US11080562B1 (en) | Key point recognition with uncertainty measurement | |
Wang et al. | A robust 6-D pose tracking approach by fusing a multi-camera tracking device and an AHRS module | |
CN113447014A (en) | Indoor mobile robot, mapping method, positioning method, and mapping positioning device | |
CN114332158A (en) | 3D real-time multi-target tracking method based on camera and laser radar fusion | |
WO2019113859A1 (en) | Machine vision-based virtual wall construction method and device, map construction method, and portable electronic device | |
CN115761693A (en) | Method for detecting vehicle location mark points and tracking and positioning vehicles based on panoramic image | |
CN102682445B (en) | Coordinate extracting algorithm of lacertilian-imitating suborder chamaeleonidae biological vision | |
WO2022226432A1 (en) | Hand gesture detection methods and systems with hand prediction | |
CN112767482B (en) | Indoor and outdoor positioning method and system with multi-sensor fusion | |
WO2022021132A1 (en) | Computer device positioning method and apparatus, computer device, and storage medium | |
CN114463832A (en) | Traffic scene sight tracking method and system based on point cloud | |
CN112484718A (en) | Edge navigation device and method based on environmental map correction | |
JPH06258028A (en) | Method and system for visually recognizing three dimensional position and attitude | |
CN112509006A (en) | Sub-map recovery fusion method and device | |
TWI780468B (en) | Method and system of robot for human following | |
WO2023130465A1 (en) | Aerial vehicle, image processing method and apparatus, and movable platform |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |