CN107260179A - Human body motion tracking method based on inertia and body-sensing sensing data quality evaluation - Google Patents
Human body motion tracking method based on inertia and body-sensing sensing data quality evaluation Download PDFInfo
- Publication number
- CN107260179A CN107260179A CN201710428293.7A CN201710428293A CN107260179A CN 107260179 A CN107260179 A CN 107260179A CN 201710428293 A CN201710428293 A CN 201710428293A CN 107260179 A CN107260179 A CN 107260179A
- Authority
- CN
- China
- Prior art keywords
- data
- sensing
- sensor
- quality evaluation
- inertial sensor
- 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
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
- A61B5/11—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor, mobility of a limb
- A61B5/1118—Determining activity level
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/68—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient
- A61B5/6801—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient specially adapted to be attached to or worn on the body surface
- A61B5/6802—Sensor mounted on worn items
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
- G06F3/011—Arrangements for interaction with the human body, e.g. for user immersion in virtual reality
Abstract
The present invention provides a kind of human body motion tracking method based on inertia and body-sensing sensing data quality evaluation, comprises the following steps:The acceleration information and positional information for obtaining skeleton artis using inertial sensor and body-sensing sensor are used as all-around exercises data;Quality evaluation will be carried out after the alignment of all-around exercises data elapsed time and space re-projection, obtain quality evaluation result;Carried out based on quality evaluation result after data fusion, choose believable all-around exercises data as new human body movement data, carry out motion tracking.The present invention effectively improves when current optics more total system carries out human motion measurement that Processing Algorithm is complicated, computationally intensive, the problem of can not be worked when target is obscured or is blocked;Human body positioning can not be carried out and receive the problem of magnetic field and metal influence by solving;Realize the mans motion simulation of low-cost and high-precision.
Description
Technical field
The invention belongs to mans motion simulation technical field, and in particular to one kind is sensed based on inertial sensor and body-sensing
The human body motion tracking method of device quality testing.
Background technology
Human Movement Tracking System is a kind of high-tech for being used for accurate measurement moving object in three-dimensional space motion situation
Equipment.Mans motion simulation (Human motion capture) is also referred to as motion capture at home, is to utilize video, inertia sensing
The movable information of device etc., in three dimensions accurate measurement human body, by the human body movement data collected, further calculating is handled
Afterwards, the motion model of human body is reconstructed in computer virtual world, so as to realize the process of the reproduction of human motion.
In November, 2010, Microsoft was proposed Kinect somatosensory game console.Kinect equipped with three camera lenses, in
Between camera lens be RGB VGA cameras, the right and left camera lens is respectively then RF transmitter and infrared C M0S cameras institute
The 3D depth inductors of composition.Utilize infrared reflective device and infrared camera capture three dimensions somatic data coordinate and speed letter
Breath, is positioned in real time.Kinect is cheap as human body motion capture equipment cost, and real-time is high, but its sphere of action has
Limit, target is easily influenceed by blocking.And inertia-type motion tracking system uses electric mechanical equipment, by the resistance for measuring gyro
The change of power, acceleration and gradient, to measure the relative motion of human body.Inertial sensor cost is low, small volume, is easy to wear
Wear, and by means of radio sensing network, the advantages of motion of human body can not be limited, but the fortune of the human body based on inertial sensor
There is also can not carry out absolute fix for motion tracking system, it is impossible to eliminates cumulative errors and is influenceed etc. to ask by magnetic field and metal object
Topic.
, the beautiful upper limb motion measurement skill that have studied based on optics and the fusion of inertia tracking data of Zhejiang University in 2014
Art, achieves certain achievement, and the human body movement data precision and reliability that this method is obtained all increase, but as a result of
The Optitrack optical tracking systems of Natural Point companies of the U.S., set more trouble and cost are high, the scope of application
Limited by cost.
Therefore, in order to solve the above problems, it is cheap simple to develop a kind of equipment, and can effectively solve the problem that by magnetic field with
And the human body motion tracking method that metal object blocks influence and causes mistake and error to occur is those skilled in the art institute
Urgent need to solve the problem.
The content of the invention
To solve the above problems, the invention discloses a kind of human body based on inertia and body-sensing sensing data quality evaluation
Motion tracking method.
In order to achieve the above object, the present invention provides following technical scheme:
A kind of human body motion tracking method based on inertia and body-sensing sensing data quality evaluation, comprises the following steps:
The acceleration information and positional information for obtaining skeleton artis using inertial sensor and body-sensing sensor are transported as whole body
Dynamic data;Quality evaluation is carried out after the alignment of all-around exercises data elapsed time and space re-projection by acquisition, quality is obtained and comments
Valency result;Based on quality evaluation result, the all-around exercises data of acquisition are carried out after data fusion, believable all-around exercises is chosen
Data carry out motion tracking as new human body movement data with this.
Further, specific steps include:
(1) acceleration information of skeleton artis, is obtained using inertial sensor and body-sensing sensor and position is believed
Breath is used as all-around exercises data:
(1.1) inertial sensor, is worn in 15 major joints of whole body, and utilizes magnetometer, accelerometer and gyro
The axis movement sensor of instrument 9 carries out Attitude Calculation, obtains joint attitude data, and joint attitude data is sent into number to host computer
According to;
(1.2) position data of 20 artis of human body, is obtained using body-sensing sensor, in the form of three dimensional space coordinate
Joint point data is recorded as, and Credibility judgement is carried out to each joint point data, believable joint point data is chosen;
(1.3) the joint attitude data and believable joint point data for, obtaining step (1.1) and step (1.2) are made
For primary motor data;
(2), the primary motor data of acquisition are aligned as the all-around exercises data elapsed time and space re-projection is laggard
Row quality evaluation, obtains quality evaluation result;
(2.1) inertial sensor and body-sensing sensor, are carried out by time alignment using interpolation method;
(2.2), obtaining space conversion matrix according to the spatial relationship between body-sensing sensor and inertial sensor will obtain
All-around exercises data carry out space conversion, and be transferred under skeleton coordinate system;
(2.3), when the new frame data of body-sensing sensor arrive, according to its confidence level, skeleton life whether is met
Reason constraint and kinematic constraint judge whether it is reliable, carry out quality evaluation;
(2.4) the reliable artis that the joint attitude data and body-sensing sensor, obtained to inertial sensor is obtained
Data carry out Kalman filtering, and filtered value is contrasted with inertial sensor observation, judge whether it exceedes setting
Filtering error threshold value;When it exceedes filtering error threshold value, auxiliary carries out quality evaluation with body-sensing sensor, whether judges it
It is credible, obtain quality evaluation result;
(3), based on quality evaluation result, the all-around exercises data of acquisition are carried out after data fusion, choose believable complete
Body exercise data carries out motion tracking as new human body movement data with this;
(3.1), the data setting one collected by inertial sensor and body-sensing sensor deposits the quality of data
Sliding window;
(3.2), when the data collected to inertial sensor and body-sensing sensor carry out quality evaluation, update respective
Sliding window;The weight of each artis is calculated according to the quantity of trust data in sliding window;
(3.3), according to the artis weight size of inertial sensor and body-sensing sensor, the joint of corresponding joint point is chosen
Point data is used as human body movement data as more believable reference data.
Further, inertial sensor is MEMS motion sensors;Body-sensing sensor is Kinect.
Further, 15 major joints are respectively in step (1.1):Head, two large arm, two forearms, waist, vertebra,
Two thighs, two shanks, two pin.
Further, 20 artis are respectively in step (1.2):Hip joint center, backbone, both shoulders center, head, a left side
Shoulder, left elbow, left wrist, left hand, right shoulder, right elbow, right wrist, the right hand, left hip joint, left knee, left ankle, left foot, right hip joint, right knee,
Right ankle, right crus of diaphragm.
Further, the specific method of step (2.3) is:The position data of 20 artis obtained from body-sensing sensor
Reliable human skeleton model is extracted as reference model, and sets up skeleton physiological bounds;According to each artis meter
The length of every section of bone is calculated, by the length of the bone calculated with the bone reference length progress pair in human skeleton model
Than if error amount is more than the error threshold of setting, illustrating that the new frame data arrived by body-sensing sensor are unreliable.
Further, the specific method of step (2.4) is:Respectively to obtaining joint attitude data conduct by inertial sensor
Inertia observation and the joint point data that is obtained by body-sensing sensor carry out Kalman filtering as body-sensing observation, are used to
Property filter value and body-sensing filter value;The error amount between inertia observation and digital filter value is calculated, and is missed with the filtering of setting
Poor threshold comparison;When error amount is less than the filtering error threshold value of setting, if meeting kinematic constraint, the inertia observation obtained
It is credible;When error amount is more than the filtering error threshold value of setting, if it is believable, inertia observation that body-sensing sensor, which obtains data,
Value is insincere, conversely, then inertia observation is credible.
Further, the specific method of step (3) is:
(3.1), one size of each joint point data setting collected by inertial sensor and body-sensing sensor is
30 sliding window, deposits the quality of data of artis;
(3.2), when a new frame observation for inertial sensor and body-sensing sensor arrives, commented according to the quality of data
Valency, judges the quality of data of each artis of present frame, updates the sliding window of inertial sensor and body-sensing sensor;If
Some joint has x frame data to be believable, then it is incredible to have 30-x frame data, and inertial sensor and body-sensing sensor are obtained
The data weighting of each artis obtained increases with the increase of the quantity of trust data in sliding window;Each artis
Weight m be expressed as:
Wherein, m is weighted value, and k is regulatory factor, and x is the believable frame number of data in sliding window;
And observation is obtained by inertial sensor and body-sensing sensor respectively and carries out weight calculation, is expressed as
mj、mi;
(3.3) m, is worked asi>=σ and mjDuring >=σ, the observation that inertial sensor and body-sensing sensor are obtained is credible, now
Using mi、mjIn a larger side Kalman filtering after joint point data as reference data, and if mi=mj, then use
Inertial sensor observation is used as reference data;
Work as mi> σ and mjDuring < σ, the error of inertial sensor observation is judged, and body-sensing sensor observation is good, by
There is larger error in the filter value after now inertial sensor and body-sensing sensor card Kalman Filtering, seen using body-sensing sensor
Measured value is used as reference data;
Work as miDuring < σ, the error of body-sensing sensor is judged, because the quality of data of inertial sensor observation is passed in body-sensing
Judged under sensor observation auxiliary, therefore mj< σ, now directly judge that inertial sensor observation is correct, using it
It is used as reference data.
It is provided by the present invention to be based on inertial sensor and body-sensing sensor mass pricer body motion tracking method,
Quality evaluation is carried out to obtaining human body movement data by inertial sensor and body-sensing sensor, according to quality evaluation result to used
Property sensor and body-sensing sensor obtain human body movement data carry out fusion treatment.Compared with prior art, (1) is effectively improved
When current optics more total system carries out human motion measurement Processing Algorithm it is complicated, computationally intensive and when target obscure or by
The problem of can not being worked when blocking;(2) determine while solving and human body can not being carried out in inertial sensor Human Movement Tracking System
The problem of position and receipts magnetic field and metal influence;(3) the inertia people based on MEMS motion sensors is realized under same framework
Body catches the human body movement data acquisition method of system and the somatosensory device based on Kinect, equipment manufacturing cost, following range,
Have both the advantage of two kind equipments in terms of positioning precision, control errors, realize the mans motion simulation of low-cost and high-precision.
Brief description of the drawings
Fig. 1, the present invention flow chart;
The quality evaluation flow chart of data is obtained in Fig. 2, the present invention to Kinect;
The quality evaluation flow chart of data is obtained in Fig. 3, the present invention to MEMS motion sensors;
The stream of human body movement data fusion treatment is carried out in Fig. 4, the present invention based on MEMS motion sensors and Kinect
Cheng Tu.
Embodiment
The technical scheme provided below with reference to specific embodiment the present invention is described in detail, it should be understood that following specific
Embodiment is only illustrative of the invention and is not intended to limit the scope of the invention.
It is as shown in Figure 1 the flow chart of the present invention, the present invention is commented to be a kind of based on inertia and body-sensing sensing data quality
The human body motion tracking method of valency, inertial sensor selection is MEMS motion sensors;The selection of body-sensing sensor is Kinect, tool
Body step includes:
(1) acceleration information of skeleton artis, is obtained using MEMS motion sensors and Kinect and position is believed
Breath is used as all-around exercises data:
(1.1) MEMS motion sensors, are worn in 15 major joints of whole body, and using magnetometer, accelerometer and
The axis movement sensor of gyroscope 9 carries out Attitude Calculation, obtains joint attitude data, and joint attitude data is sent to host computer
Data;Wherein 15 major joints are respectively:Head, two large arm, two forearms, waist, vertebra, two thighs, two shanks, two
Pin;
(1.2) position data of 20 artis of human body, is obtained using Kinect, is recorded in the form of three dimensional space coordinate
For joint point data, and Credibility judgement is carried out to each joint point data, choose believable joint point data;Such as Fig. 2, when
When the state of artis is in TRACKED, then it is assumed that joint point data is tentatively credible, then judges whether it meets physiology about
Beam and kinematic constraint, if be satisfied by, then it is assumed that credible;Wherein 20 artis are respectively:It is hip joint center, backbone, double
Shoulder center, head, left shoulder, left elbow, left wrist, left hand, right shoulder, right elbow, right wrist, the right hand, left hip joint, left knee, left ankle, left foot,
Right hip joint, right knee, right ankle, right crus of diaphragm;
(1.3) the joint attitude data and believable joint point data for, obtaining step (1.1) and step (1.2) are made
For primary motor data;
(2), the primary motor data of acquisition are aligned as the all-around exercises data elapsed time and space re-projection is laggard
Row quality evaluation, obtains quality evaluation result;
(2.1) MEMS motion sensors and Kinect, are carried out by time alignment using interpolation method;
According to MEMS motion sensors and the frequency of Kinect gathered data, timeslice is chosen for 1 second;MEMS is transported
Dynamic sensor actual acquisition speed ratio Kinect is low, in order that the data of MEMS motion sensors formed same Kinect it is the same when
Between be spaced, it is assumed that the uniform motion of constant angular speed is done in the time interval of the frame of MEMS motion sensors two, by interpolation method meter
The MEMS motion sensor angles obtained after calculation, now the time interval of MEMS motion sensor datas is consistent with Kinect;
(2.2), obtaining space conversion matrix according to the spatial relationship between Kinect and MEMS motion sensors will obtain
All-around exercises data carry out space conversion, and be transferred under skeleton coordinate system;
(2.3), the position data from Kinect 20 artis obtained extracts reliable human skeleton model conduct
Reference model, and set up skeleton physiological bounds;The length of every section of bone is calculated according to each artis, by what is calculated
The length of bone is contrasted with the bone reference length in human skeleton model, if error amount is more than the error threshold of setting,
Then illustrate that the new frame data arrived by body-sensing sensor are unreliable;
Due to the kinematic constraint of human body itself, the angle that the artis of human body different parts can rotate also has respective
All artis are carried out Eulerian angles calculating according to the skeleton data got, judge the Euler of each artis by span
Whether angle is in corresponding scope;If Eulerian angles go beyond the scope more than error threshold, then it is assumed that the joint points in this frame
According to insincere;Following table is the Eulerian angles span of human body parts artis:
(2.4) the reliable artis that the joint attitude data and body-sensing sensor, obtained to inertial sensor is obtained
Data carry out Kalman filtering, and filtered value is contrasted with inertial sensor observation, judge whether it exceedes setting
Filtering error threshold value;When it exceedes filtering error threshold value, auxiliary carries out quality evaluation with body-sensing sensor, whether judges it
It is credible, obtain quality evaluation result;
When two equipment a data wherein side error when, due to the complementarity of information, cause MEMS motion sensors and
Kinect filtering error value all produces large change.One filtering error threshold is set for the filtering error of inertial sensor to this
Value ε, when the filtering error value of MEMS motion sensors is less than ε, then whether the data for judging now meet the fortune of human synovial
Moving constraint, if meeting kinematic constraint, then it represents that the data of MEMS motion sensors are credible;When the filtering of MEMS motion sensors
When error amount is more than ε, illustrate that one of MEMS motion sensors and Kinect make a mistake, cause filtering error to become big;If
The Kinect quality of data is credible, then it is assumed that is due to that filtering error becomes big caused by MEMS motion sensors, otherwise thinks
The data of MEMS motion sensors are believable, specific method such as Fig. 3:
Respectively inertia observation S is used as to obtaining joint attitude data by MEMS motion sensorsobserveAnd by
The joint point data that Kinect is obtained is used as body-sensing observation KobserveKalman filtering is carried out, digital filter value S is obtainedkalman
With body-sensing filter value Kkalman;Calculate SobserveWith SkalmanBetween error amount Dsensor, and with the filtering error threshold epsilon of setting
Contrast;Work as DsensorDuring less than ε, if meeting kinematic constraint, the S obtainedobserveIt is credible;Work as DsensorDuring more than ε, if Kinect
It is believable, then S to obtain dataobserveIt is insincere, conversely, then SobserveIt is credible;Quality evaluation result is obtained with this;
(3), based on quality evaluation result, the all-around exercises data of acquisition are carried out after data fusion, choose believable complete
Body exercise data carries out motion tracking as new human body movement data with this;
(3.1), one size of each joint point data setting collected by MEMS motion sensors and Kinect is
30 sliding window, deposits the quality of data of artis;
(3.2), when a new frame observation for MEMS motion sensors and Kinect arrives, commented according to the quality of data
Valency, judges the quality of data of each artis of present frame, updates MEMS motion sensors and Kinect sliding window;If
Some joint has x frame data to be believable, then it is incredible to have 30-x frame data, and MEMS motion sensors and Kinect are obtained
The data weighting of each artis obtained increases with the increase of the quantity of trust data in sliding window;Each artis
Weight m be expressed as:
Wherein, m is weighted value, and k is regulatory factor, and x is the believable frame number of data in sliding window;
Such as Fig. 4, and pass through MEMS motion sensors and Kinect acquisition observation progress weight calculations respectively, respectively
It is expressed as mj、mi;
(3.3) m, is worked asi>=σ and mjDuring >=σ, the observation that MEMS motion sensors and Kinect are obtained is credible, now
Using mi、mjIn a larger side Kalman filtering after joint point data as reference data, and if mi=mj, then use
The observation of MEMS motion sensors is used as reference data;
Work as mi> σ and mjDuring < σ, the observation error of MEMS motion sensors is judged, and Kinect observations are good,
Because the filter value after now MEMS motion sensors and Kinect Kalman filterings has larger error, using Kinect sight
Measured value is used as reference data;
Work as miDuring < σ, Kinect errors are judged, because the observation quality of data of MEMS motion sensors is in Kinect
Observation auxiliary under judge, therefore mj< σ, now directly judge that the observation of MEMS motion sensors is correct, adopt
Reference data is used as with it.
It is last it should be noted that the above embodiments are merely illustrative of the technical solutions of the present invention and non-limiting technical side
Case, it will be understood by those within the art that, those modify or equivalent substitution to technical scheme, and
The objective and scope of the technical program are not departed from, all should be covered among scope of the presently claimed invention.
Claims (8)
1. a kind of human body motion tracking method based on inertia and body-sensing sensing data quality evaluation, it is characterised in that:Including
Following steps:The acceleration information and positional information for obtaining skeleton artis using inertial sensor and body-sensing sensor are made
For all-around exercises data;Quality evaluation is carried out after the alignment of all-around exercises data elapsed time and space re-projection by acquisition, is obtained
Obtain quality evaluation result;Based on quality evaluation result, the all-around exercises data of acquisition are carried out after data fusion, chosen believable
All-around exercises data carry out motion tracking as new human body movement data with this.
2. a kind of mans motion simulation side based on inertia and body-sensing sensing data quality evaluation according to claim 1
Method, it is characterised in that:Specific steps include:
(1), the acceleration information and positional information for obtaining skeleton artis using inertial sensor and body-sensing sensor are made
For all-around exercises data:
(1.1) inertial sensor, is worn in 15 major joints of whole body, and utilizes magnetometer, accelerometer and the axle of gyroscope 9
Motion sensor carries out Attitude Calculation, obtains joint attitude data, and joint attitude data is sent into data to host computer;
(1.2) position data of 20 artis of human body, is obtained using body-sensing sensor, is recorded in the form of three dimensional space coordinate
For joint point data, and Credibility judgement is carried out to each joint point data, choose believable joint point data;
(1.3) the joint attitude data that, obtains step (1.1) with step (1.2) and believable joint point data are as first
Level exercise data;
(2), using the primary motor data of acquisition as the all-around exercises data elapsed time be aligned and space re-projection after carry out matter
Amount is evaluated, and obtains quality evaluation result;
(2.1) inertial sensor and body-sensing sensor, are carried out by time alignment using interpolation method;
(2.2) space conversion matrix, is obtained by the complete of acquisition according to the spatial relationship between body-sensing sensor and inertial sensor
Body exercise data carries out space conversion, and is transferred under skeleton coordinate system;
(2.3), when the new frame data of body-sensing sensor arrive, according to its confidence level, skeleton physiology whether is met about
Beam and kinematic constraint judge whether it is reliable, carry out quality evaluation;
(2.4) the reliable joint point data that the joint attitude data and body-sensing sensor, obtained to inertial sensor is obtained
Kalman filtering is carried out, filtered value is contrasted with inertial sensor observation, its filter for whether exceeding setting is judged
Wave error threshold value;When it exceedes filtering error threshold value, auxiliary carries out quality evaluation with body-sensing sensor, judges that it whether may be used
Letter, obtains quality evaluation result;
(3), based on quality evaluation result, the all-around exercises data of acquisition are carried out after data fusion, believable whole body fortune is chosen
Dynamic data carry out motion tracking as new human body movement data with this;
(3.1), the slip of the one storage quality of data of data setting collected by inertial sensor and body-sensing sensor
Window;
(3.2), when the data collected to inertial sensor and body-sensing sensor carry out quality evaluation, respective cunning is updated
Dynamic window;The weight of each artis is calculated according to the quantity of trust data in sliding window;
(3.3), according to the artis weight size of inertial sensor and body-sensing sensor, the joint points of corresponding joint point are chosen
Human body movement data is used as according to as more believable reference data.
3. a kind of mans motion simulation side based on inertia and body-sensing sensing data quality evaluation according to claim 2
Method, it is characterised in that:The inertial sensor is MEMS motion sensors;The body-sensing sensor is Kinect.
4. a kind of mans motion simulation side based on inertia and body-sensing sensing data quality evaluation according to claim 2
Method, it is characterised in that:15 major joints are respectively in the step (1.1):Head, two large arm, two forearms, waist, vertebra,
Two thighs, two shanks, two pin.
5. a kind of mans motion simulation side based on inertia and body-sensing sensing data quality evaluation according to claim 2
Method, it is characterised in that:20 artis are respectively in the step (1.2):Hip joint center, backbone, both shoulders center, head,
Left shoulder, left elbow, left wrist, left hand, right shoulder, right elbow, right wrist, the right hand, left hip joint, left knee, left ankle, left foot, right hip joint, the right side
Knee, right ankle, right crus of diaphragm.
6. a kind of mans motion simulation side based on inertia and body-sensing sensing data quality evaluation according to claim 2
Method, it is characterised in that:The specific method of the step (2.3) is:The positional number of 20 artis obtained from body-sensing sensor
According to extracting reliable human skeleton model as reference model, and set up skeleton physiological bounds;According to each artis
The length of every section of bone is calculated, by the length of the bone calculated with the bone reference length progress pair in human skeleton model
Than if error amount is more than the error threshold of setting, illustrating that the new frame data arrived by body-sensing sensor are unreliable.
7. a kind of mans motion simulation side based on inertia and body-sensing sensing data quality evaluation according to claim 2
Method, it is characterised in that:The specific method of the step (2.4) is:Make respectively to obtaining joint attitude data by inertial sensor
The joint point data obtained for inertia observation and by body-sensing sensor carries out Kalman filtering as body-sensing observation, obtains
Digital filter value and body-sensing filter value;Calculate the error amount between inertia observation and digital filter value, and with the filtering of setting
Error threshold is contrasted;When error amount is less than the filtering error threshold value of setting, if meeting kinematic constraint, the inertia observation obtained
Value is credible;When error amount is more than the filtering error threshold value of setting, if it is believable that body-sensing sensor, which obtains data, inertia is seen
Measured value is insincere, conversely, then inertia observation is credible.
8. a kind of mans motion simulation side based on inertia and body-sensing sensing data quality evaluation according to claim 2
Method, it is characterised in that:The specific method of the step (3) is:
(3.1), each joint point data collected by inertial sensor and body-sensing sensor sets a size to be 30
Sliding window, deposits the quality of data of artis;
(3.2), when a new frame observation for inertial sensor and body-sensing sensor arrives, according to quality testing, sentence
The quality of data for each artis for present frame of breaking, updates the sliding window of inertial sensor and body-sensing sensor;If a certain
Individual joint has x frame data to be believable, then have 30-x frame data be it is incredible, what inertial sensor and body-sensing sensor were obtained
The data weighting of each artis increases with the increase of the quantity of trust data in sliding window;The power of each artis
Weight m is expressed as:
<mrow>
<mi>m</mi>
<mo>=</mo>
<mi>exp</mi>
<mrow>
<mo>(</mo>
<mo>-</mo>
<mi>k</mi>
<mfrac>
<mrow>
<mn>30</mn>
<mo>-</mo>
<mi>x</mi>
</mrow>
<mn>30</mn>
</mfrac>
<mo>)</mo>
</mrow>
</mrow>
Wherein, m is weighted value, and k is regulatory factor, and x is the believable frame number of data in sliding window;
And observation is obtained by inertial sensor and body-sensing sensor respectively and carries out weight calculation, m is expressed asj、
mi;
(3.3) m, is worked asi>=σ and mjDuring >=σ, the observation that inertial sensor and body-sensing sensor are obtained is credible, now uses
mi、mjIn a larger side Kalman filtering after joint point data as reference data, and if mi=mj, then using inertia
Sensor observation is used as reference data;
Work as mi> σ and mjDuring < σ, the error of inertial sensor observation is judged, and body-sensing sensor observation is good, due to this
When inertial sensor and body-sensing sensor card Kalman Filtering after filter value there is larger error, using body-sensing sensor observation
It is used as reference data;
Work as miDuring < σ, the error of body-sensing sensor is judged, because the quality of data of inertial sensor observation is in body-sensing sensor
Judged under observation auxiliary, therefore mj< σ, now directly judge that inertial sensor observation is correct, using its conduct
Reference data.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710428293.7A CN107260179A (en) | 2017-06-08 | 2017-06-08 | Human body motion tracking method based on inertia and body-sensing sensing data quality evaluation |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710428293.7A CN107260179A (en) | 2017-06-08 | 2017-06-08 | Human body motion tracking method based on inertia and body-sensing sensing data quality evaluation |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107260179A true CN107260179A (en) | 2017-10-20 |
Family
ID=60066979
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710428293.7A Pending CN107260179A (en) | 2017-06-08 | 2017-06-08 | Human body motion tracking method based on inertia and body-sensing sensing data quality evaluation |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107260179A (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109540132A (en) * | 2018-11-22 | 2019-03-29 | 中国矿业大学 | Localization method of the movable equipment based on sensor fusion on the person |
CN109991889A (en) * | 2017-12-29 | 2019-07-09 | 深圳光启尖端技术有限责任公司 | The control device and walking robot of walking robot |
CN110132257A (en) * | 2019-05-15 | 2019-08-16 | 吉林大学 | Human body behavior prediction method based on Fusion |
CN110852162A (en) * | 2019-09-29 | 2020-02-28 | 深圳云天励飞技术有限公司 | Human body integrity data labeling method and device and terminal equipment |
CN111506199A (en) * | 2020-05-06 | 2020-08-07 | 北京理工大学 | Kinect-based high-precision unmarked whole-body motion tracking system |
CN111887856A (en) * | 2020-02-25 | 2020-11-06 | 哈尔滨工业大学 | Inertial sensor-based real-time calculation method for position-drunkenness-resistant joint angle |
CN112330714A (en) * | 2020-09-29 | 2021-02-05 | 深圳大学 | Pedestrian tracking method and device, electronic equipment and storage medium |
CN113556600A (en) * | 2021-07-13 | 2021-10-26 | 广州虎牙科技有限公司 | Drive control method and device based on time sequence information, electronic equipment and readable storage medium |
CN114271814A (en) * | 2021-12-24 | 2022-04-05 | 安徽大学 | Kinect-based rehabilitation training and evaluation method and system for stroke patient |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102175244A (en) * | 2011-03-16 | 2011-09-07 | 公安部沈阳消防研究所 | In-building person positioning system and positioning method thereof based on inertial sensor |
CN103399637A (en) * | 2013-07-31 | 2013-11-20 | 西北师范大学 | Man-computer interaction method for intelligent human skeleton tracking control robot on basis of kinect |
CN103921266A (en) * | 2014-04-15 | 2014-07-16 | 哈尔滨工程大学 | Method for somatosensory control over snow and robot on basis of Kinect |
CN103933722A (en) * | 2014-02-28 | 2014-07-23 | 永康市坤卓科技有限公司 | Bodybuilding dumbbell exercise detection device and method |
CN104658012A (en) * | 2015-03-05 | 2015-05-27 | 第二炮兵工程设计研究院 | Motion capture method based on inertia and optical measurement fusion |
CN105068654A (en) * | 2015-08-14 | 2015-11-18 | 济南中景电子科技有限公司 | Motion capturing system and method based on CAN bus and inertial sensor |
CN105551059A (en) * | 2015-12-08 | 2016-05-04 | 国网山西省电力公司技能培训中心 | Power transformation simulation human body motion capturing method based on optical and inertial body feeling data fusion |
-
2017
- 2017-06-08 CN CN201710428293.7A patent/CN107260179A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102175244A (en) * | 2011-03-16 | 2011-09-07 | 公安部沈阳消防研究所 | In-building person positioning system and positioning method thereof based on inertial sensor |
CN103399637A (en) * | 2013-07-31 | 2013-11-20 | 西北师范大学 | Man-computer interaction method for intelligent human skeleton tracking control robot on basis of kinect |
CN103933722A (en) * | 2014-02-28 | 2014-07-23 | 永康市坤卓科技有限公司 | Bodybuilding dumbbell exercise detection device and method |
CN103921266A (en) * | 2014-04-15 | 2014-07-16 | 哈尔滨工程大学 | Method for somatosensory control over snow and robot on basis of Kinect |
CN104658012A (en) * | 2015-03-05 | 2015-05-27 | 第二炮兵工程设计研究院 | Motion capture method based on inertia and optical measurement fusion |
CN105068654A (en) * | 2015-08-14 | 2015-11-18 | 济南中景电子科技有限公司 | Motion capturing system and method based on CAN bus and inertial sensor |
CN105551059A (en) * | 2015-12-08 | 2016-05-04 | 国网山西省电力公司技能培训中心 | Power transformation simulation human body motion capturing method based on optical and inertial body feeling data fusion |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109991889A (en) * | 2017-12-29 | 2019-07-09 | 深圳光启尖端技术有限责任公司 | The control device and walking robot of walking robot |
CN109540132A (en) * | 2018-11-22 | 2019-03-29 | 中国矿业大学 | Localization method of the movable equipment based on sensor fusion on the person |
CN110132257A (en) * | 2019-05-15 | 2019-08-16 | 吉林大学 | Human body behavior prediction method based on Fusion |
CN110852162B (en) * | 2019-09-29 | 2020-10-23 | 深圳云天励飞技术有限公司 | Human body integrity data labeling method and device and terminal equipment |
CN110852162A (en) * | 2019-09-29 | 2020-02-28 | 深圳云天励飞技术有限公司 | Human body integrity data labeling method and device and terminal equipment |
CN111887856A (en) * | 2020-02-25 | 2020-11-06 | 哈尔滨工业大学 | Inertial sensor-based real-time calculation method for position-drunkenness-resistant joint angle |
CN111887856B (en) * | 2020-02-25 | 2022-09-16 | 哈尔滨工业大学 | Inertial sensor-based real-time calculation method for position-drunkenness-resistant joint angle |
CN111506199A (en) * | 2020-05-06 | 2020-08-07 | 北京理工大学 | Kinect-based high-precision unmarked whole-body motion tracking system |
CN112330714A (en) * | 2020-09-29 | 2021-02-05 | 深圳大学 | Pedestrian tracking method and device, electronic equipment and storage medium |
CN112330714B (en) * | 2020-09-29 | 2024-01-09 | 深圳大学 | Pedestrian tracking method and device, electronic equipment and storage medium |
CN113556600A (en) * | 2021-07-13 | 2021-10-26 | 广州虎牙科技有限公司 | Drive control method and device based on time sequence information, electronic equipment and readable storage medium |
CN113556600B (en) * | 2021-07-13 | 2023-08-18 | 广州虎牙科技有限公司 | Drive control method and device based on time sequence information, electronic equipment and readable storage medium |
CN114271814A (en) * | 2021-12-24 | 2022-04-05 | 安徽大学 | Kinect-based rehabilitation training and evaluation method and system for stroke patient |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107260179A (en) | Human body motion tracking method based on inertia and body-sensing sensing data quality evaluation | |
CN104658012B (en) | Motion capture method based on inertia and optical measurement fusion | |
KR102190932B1 (en) | Method of determining joint stress from sensor data | |
Chiari et al. | Human movement analysis using stereophotogrammetry: Part 2: Instrumental errors | |
Lin et al. | Human pose recovery using wireless inertial measurement units | |
US8743054B2 (en) | Graphical representations | |
Peppoloni et al. | A novel 7 degrees of freedom model for upper limb kinematic reconstruction based on wearable sensors | |
CN109141433A (en) | A kind of robot indoor locating system and localization method | |
CN105551059A (en) | Power transformation simulation human body motion capturing method based on optical and inertial body feeling data fusion | |
CN109284006B (en) | Human motion capturing device and method | |
WO2009007917A2 (en) | Object motion capturing system and method | |
Chen et al. | Real-time human motion capture driven by a wireless sensor network | |
CN110609621B (en) | Gesture calibration method and human motion capture system based on microsensor | |
CN106256394A (en) | The training devices of mixing motion capture and system | |
CN208751577U (en) | A kind of robot indoor locating system | |
CN113793360B (en) | Three-dimensional human body reconstruction method based on inertial sensing technology | |
CN103340632A (en) | Human joint angle measuring method based on feature point space position | |
Salehi et al. | Body-IMU autocalibration for inertial hip and knee joint tracking | |
CN108734762B (en) | Motion trail simulation method and system | |
CN109767470A (en) | A kind of tracking system initial method and terminal device | |
CN108534772A (en) | Attitude angle acquisition methods and device | |
CN109453505B (en) | Multi-joint tracking method based on wearable device | |
CN115919250A (en) | Human dynamic joint angle measuring system | |
CN108013880A (en) | A kind of instantaneous aroused in interest measuring method of human elbow anterior flexion and rear stretching around instantaneous aroused in interest movement | |
CN106886288A (en) | A kind of attitude dynamic method for catching 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 |