CN104866846A - Action recognition device and method - Google Patents
Action recognition device and method Download PDFInfo
- Publication number
- CN104866846A CN104866846A CN201510321587.0A CN201510321587A CN104866846A CN 104866846 A CN104866846 A CN 104866846A CN 201510321587 A CN201510321587 A CN 201510321587A CN 104866846 A CN104866846 A CN 104866846A
- Authority
- CN
- China
- Prior art keywords
- action
- data
- primary
- sensor
- calculation element
- 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.)
- Withdrawn
Links
Classifications
-
- 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
- G06V40/20—Movements or behaviour, e.g. gesture recognition
- G06V40/23—Recognition of whole body movements, e.g. for sport training
Abstract
The invention discloses an action recognition device and method, and relates to the field of mode recognition. One end of an image sensing device is connected with an action sensor, the other end of the image sensing device is connected with one end of a primary action detection device, the other end of the primary action detection device is connected with one end of an efficient action data extracting device, the other end of the effective action data extracting device is connected with one end of a data storage device, the other end of the data storage device is connected with one end of a communication device, and the other end of the communication device is connected with one end of an action computation device. A primary action recognition module and a data storage module are added in a data acquiring device to realize primary screening and caching of motion data at a data acquiring device end, even if a little noise points are introduced into the data acquiring process, a value of median of fitting data in a sliding window is hard to influence, the noise influence can be effectively avoided while cutting action data, and the accuracy rate of an action recognition result is improved.
Description
Technical field:
The present invention relates to area of pattern recognition, be specifically related to a kind of action recognition apparatus and method.
Background technology:
Along with the development of Computer Applied Technology, action recognition has been widely used in the fields such as physical culture, game, film, surgery simulation or motor skill training.Such as the data collector comprising sensor is arranged on club or gloves, user is carrying out such as golf, during the ball game such as shuttlecock, data collector collects exercise data and the exercise data collected is sent to action calculation element, this action calculation element can be arranged at such as mobile phone, in the intelligent terminals such as panel computer, after action calculation element carries out computational analysis to the exercise data that data collector sends, the position and attitude information etc. during user movement can be got, thus carry out data sharing for user, obtain action director etc. and basis is provided.
But, carried out the real-time Transmission of exercise data by wireless channel between usual data collector and action calculation element, the algorithm that action calculation element adopts when calculating is integral algorithm based on acceleration, and therefore the precision of the higher position that calculates of same noise prerequisite down-sampling rate and attitude information is also higher.If the exercise data that under adopting high sampling rate situation, large quantity sensor collects is huge concerning pressure the transmission of wireless channel, except producing a large amount of radio power consumption, the ability to bear of wireless channel even may be exceeded.
Summary of the invention:
The object of this invention is to provide a kind of action recognition apparatus and method, its structure is simple, novelty reasonable in design, by adding primary action identification module and data memory module in data collector, preliminary screening and the buffer memory of exercise data is realized at data collector end, get rid of those data to the useless static state of action recognition, then action calculation element is sent to carry out concrete action recognition the exercise data of buffer memory by the mode of active push or response request, even if introduce a small amount of noise spot in the process of image data, also the value of the median having influence on fitting data in moving window is difficult to, effectively can avoid the impact of the noise when splitting action data, improve the accuracy rate of action recognition result.
In order to solve the problem existing for background technology, the present invention is by the following technical solutions: it comprises Image sensor apparatus 1, action sensor 2, primary action pick-up unit 3, effective action data extraction device 4, data storage device 5, communication device 6, action calculation element 7 and result decision maker 8, one end of Image sensor apparatus 1 is connected with action sensor 2, the other end of Image sensor apparatus 1 is connected with one end of primary action pick-up unit 3, the other end of primary action pick-up unit 3 is connected with one end of effective action data extraction device 4, the other end of effective action data extraction device 4 is connected with one end of data storage device 5, the other end of data storage device 5 is connected with one end of communication device 6, the other end of communication device 6 is connected with one end of action calculation element 7, the other end of action calculation element 7 is connected with result decision maker 8.
Described action sensor 2 can adopt one in acceleration transducer, gyroscope, magnetic field sensor, compass, 3-axis acceleration sensor, three-axis gyroscope, triaxial magnetic field sensor and multiple sensors.
Method of work of the present invention is: action object produces action, Image sensor apparatus 1 senses signal, action data is gathered by action sensor 2, for action data digital simulation data, fitting data and the first threshold preset are compared, when fitting data is greater than first threshold, primary action pick-up unit 3 starts operation of recording data, acts of determination is in initial state, when fitting data is less than first threshold, acts of determination remains static, if rest time is greater than scheduled duration, then stop operation of recording data, scheduled duration can rule of thumb set, effective action data extraction device 4 extracted valid data when primary action pick-up unit 3 starts operation of recording data, valid data are stored in data storage device 5, then data send action calculation element 7 to by communication device 6, action calculation element 7 calculates result and shows in result decision maker 8.
The present invention has following beneficial effect: its structure is simple, novelty reasonable in design, by adding primary action identification module and data memory module in data collector, preliminary screening and the buffer memory of exercise data is realized at data collector end, get rid of those data to the useless static state of action recognition, then action calculation element is sent to carry out concrete action recognition the exercise data of buffer memory by the mode of active push or response request, even if introduce a small amount of noise spot in the process of image data, also the value of the median having influence on fitting data in moving window is difficult to, effectively can avoid the impact of the noise when splitting action data, improve the accuracy rate of action recognition result.
Accompanying drawing illustrates:
Fig. 1 is structural representation of the present invention.
Embodiment:
With reference to Fig. 1, this embodiment takes following technical scheme: it comprises Image sensor apparatus 1, action sensor 2, primary action pick-up unit 3, effective action data extraction device 4, data storage device 5, communication device 6, action calculation element 7 and result decision maker 8, one end of Image sensor apparatus 1 is connected with action sensor 2, the other end of Image sensor apparatus 1 is connected with one end of primary action pick-up unit 3, the other end of primary action pick-up unit 3 is connected with one end of effective action data extraction device 4, the other end of effective action data extraction device 4 is connected with one end of data storage device 5, the other end of data storage device 5 is connected with one end of communication device 6, the other end of communication device 6 is connected with one end of action calculation element 7, the other end of action calculation element 7 is connected with result decision maker 8.
Described action sensor 2 can adopt one in acceleration transducer, gyroscope, magnetic field sensor, compass, 3-axis acceleration sensor, three-axis gyroscope, triaxial magnetic field sensor and multiple sensors.
The method of work of this embodiment is: action object produces action, Image sensor apparatus 1 senses signal, action data is gathered by action sensor 2, for action data digital simulation data, fitting data and the first threshold preset are compared, when fitting data is greater than first threshold, primary action pick-up unit 3 starts operation of recording data, acts of determination is in initial state, when fitting data is less than first threshold, acts of determination remains static, if rest time is greater than scheduled duration, then stop operation of recording data, scheduled duration can rule of thumb set, effective action data extraction device 4 extracted valid data when primary action pick-up unit 3 starts operation of recording data, valid data are stored in data storage device 5, then data send action calculation element 7 to by communication device 6, action calculation element 7 calculates result and shows in result decision maker 8.
This embodiment has following beneficial effect: its structure is simple, novelty reasonable in design, by adding primary action identification module and data memory module in data collector, preliminary screening and the buffer memory of exercise data is realized at data collector end, get rid of those data to the useless static state of action recognition, then action calculation element is sent to carry out concrete action recognition the exercise data of buffer memory by the mode of active push or response request, even if introduce a small amount of noise spot in the process of image data, also the value of the median having influence on fitting data in moving window is difficult to, effectively can avoid the impact of the noise when splitting action data, improve the accuracy rate of action recognition result.
Claims (3)
1. action recognition apparatus and method, it is characterized in that it comprises Image sensor apparatus (1), action sensor (2), primary action pick-up unit (3), effective action data extraction device (4), data storage device (5), communication device (6), action calculation element (7) and result decision maker (8), one end of Image sensor apparatus (1) is connected with action sensor (2), the other end of Image sensor apparatus (1) is connected with one end of primary action pick-up unit (3), the other end of primary action pick-up unit (3) is connected with one end of effective action data extraction device (4), the other end of effective action data extraction device (4) is connected with one end of data storage device (5), the other end of data storage device (5) is connected with one end of communication device (6), the other end of communication device (6) is connected with one end of action calculation element (7), the other end of action calculation element (7) is connected with result decision maker (8).
2. a kind of action recognition apparatus and method according to claim 1, is characterized in that described action sensor (2) can adopt one in acceleration transducer, gyroscope, magnetic field sensor, compass, 3-axis acceleration sensor, three-axis gyroscope, triaxial magnetic field sensor and multiple sensors.
3. a kind of action recognition apparatus and method according to claim 1, it is characterized in that action object produces action, Image sensor apparatus (1) senses signal, action data is gathered by action sensor (2), for action data digital simulation data, fitting data and the first threshold preset are compared, when fitting data is greater than first threshold, primary action pick-up unit (3) starts operation of recording data, acts of determination is in initial state, when fitting data is less than first threshold, acts of determination remains static, if rest time is greater than scheduled duration, then stop operation of recording data, scheduled duration can rule of thumb set, effective action data extraction device (4) extracted valid data during primary action pick-up unit (3) beginning operation of recording data, valid data are stored in data storage device (5), then data send action calculation element (7) to by communication device (6), action calculation element (7) calculates result and shows in result decision maker (8).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510321587.0A CN104866846A (en) | 2015-06-12 | 2015-06-12 | Action recognition device and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510321587.0A CN104866846A (en) | 2015-06-12 | 2015-06-12 | Action recognition device and method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104866846A true CN104866846A (en) | 2015-08-26 |
Family
ID=53912667
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510321587.0A Withdrawn CN104866846A (en) | 2015-06-12 | 2015-06-12 | Action recognition device and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104866846A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110052000A (en) * | 2019-04-12 | 2019-07-26 | 漳州泰里斯体育器材有限公司 | A kind of identifying processing method and system of combat sports state |
-
2015
- 2015-06-12 CN CN201510321587.0A patent/CN104866846A/en not_active Withdrawn
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110052000A (en) * | 2019-04-12 | 2019-07-26 | 漳州泰里斯体育器材有限公司 | A kind of identifying processing method and system of combat sports state |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103076884B (en) | Data acquisition method and data acquisition device for motion recognition, and motion recognition system | |
| TWI786313B (en) | Method, device, storage medium, and apparatus of tracking target | |
| CN102221369B (en) | Gesture recognizing method and device of ball game and gesture auxiliary device | |
| CN104731307B (en) | A kind of body-sensing action identification method and human-computer interaction device | |
| US8725452B2 (en) | Method of confirming motion parameters apparatus for the same, and motion assisting device | |
| CN107635204B (en) | Indoor fusion positioning method and device assisted by exercise behaviors and storage medium | |
| CN103460221B (en) | For User Activity is classified using the combination of likelihood function value in the mobile device system, method and apparatus | |
| US10067157B2 (en) | Methods and systems for sensor-based vehicle acceleration determination | |
| CN102667672A (en) | Acceleration motion identify method and system thereof | |
| CN110113116B (en) | Human behavior identification method based on WIFI channel information | |
| CN104197929B (en) | Localization method, device and system based on geomagnetism and WIFI | |
| CN102671356A (en) | Swing analysis device, program, and swing analysis method | |
| CN104225891A (en) | Motion analysis method and motion analysis device | |
| CN102553194A (en) | Swing analyzing apparatus | |
| CN103699795A (en) | Exercise behavior identification method and device and exercise intensity monitoring system | |
| CN113743177A (en) | Key point detection method, system, intelligent terminal and storage medium | |
| CN101879376A (en) | Realization method for gyro sensor in interactive games | |
| CN107454970A (en) | A kind of System and method for of movement locus collection and analysis based on ball game | |
| CN104463087A (en) | Action identification device and method | |
| CN112995757B (en) | Video clipping method and device | |
| CN104866846A (en) | Action recognition device and method | |
| CN107392106B (en) | Human activity endpoint detection method based on double thresholds | |
| CN111435083A (en) | Pedestrian track calculation method, navigation method and device, handheld terminal and medium | |
| CN110909687B (en) | Action feature validity determination method, computer storage medium, and electronic device | |
| CN210078765U (en) | Motion capture recognition and evaluation device based on wearable sensor |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| EXSB | Decision made by sipo to initiate substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20150826 |
|
| WW01 | Invention patent application withdrawn after publication |