CN110275614A - A kind of non-contact gesture identification device and its method - Google Patents
A kind of non-contact gesture identification device and its method Download PDFInfo
- Publication number
- CN110275614A CN110275614A CN201910460084.XA CN201910460084A CN110275614A CN 110275614 A CN110275614 A CN 110275614A CN 201910460084 A CN201910460084 A CN 201910460084A CN 110275614 A CN110275614 A CN 110275614A
- Authority
- CN
- China
- Prior art keywords
- gesture
- chip microcontroller
- lath
- finger
- decision tree
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- 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
- G06F3/017—Gesture based interaction, e.g. based on a set of recognized hand gestures
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Human Computer Interaction (AREA)
- User Interface Of Digital Computer (AREA)
Abstract
The present invention discloses a kind of non-contact gesture identification device and its method, it includes FDC2214 capacitance sensing chip, single-side brass plate and single-chip microcontroller, single-side brass plate and gesture to be measured constitute switched capacitances, FDC2214 capacitance sensing chip is electrically connected with the copper sheet of single-side brass plate, and detect the switched capacitances value for obtaining single-side brass plate and gesture to be measured composition, FDC2214 capacitance sensing chip output accesses single-chip microcontroller, and single-chip microcontroller reads the frequency data for the reflection switched capacitances value that FDC2214 capacitance sensing chip measures;Single-chip microcontroller has the memory for saving frequency data, and single-chip microcontroller passes through the categorised decision tree of training construction complete difference gesture using frequency data and carries out gesture classification and output category result to the frequency data acquired in real time using categorised decision tree.The present invention is applicable in the gesture identification of different hand sizes, and recognition accuracy is high.
Description
Technical field
The present invention relates to technical field of hand gesture recognition more particularly to a kind of non-contact gesture identification device and its methods.
Background technique
As shown in Figure 1, FDC2214 is that a of Texas Instrument's release is passed based on LC resonance circuit theory, towards condenser type
The antinoise and electromagnetic interference, high-resolution, high speed, the detection sensor of multichannel for feeling solution, may be implemented sensing area
Interior contactless sensing.FDC2214 testing principle is as shown in Figure 1.
FDC2214 provides 4 sense channel CH0~CH3, and the input terminal of each sense channel connects an inductance and electricity
Hold, forms lc circuit.The sensor ends of measured capacitance (grey identification division in Fig. 1) are connected with lc circuit, will generate a vibration
Frequency is swung, measured capacitance values can be calculated according to the frequency of oscillation value.
Series monolithic is 32 based on AMR Cortex-M4 core that the Pu En Zhi company (former Freescale company) releases
Microcontroller, the series monolithic have many advantages, such as that low-power consumption, peripheral module are abundant.The present invention is using in series
MDN512ZVLQ10 single-chip microcontroller (calls single-chip microcontroller in the following text), has 512KB flash memory, 128KB random access memory in the model single chip microcomputer
It is empty to can satisfy storage required when storage and the operational decisions tree construction algorithm of training stage a large amount of gesture datas by SRAM
Between.Meanwhile single-chip microcontroller has multiple communication interface, such as I2C communication interface, ethernet mac controller, USB communication interface, SD master
Machine controller, UART communication interface etc., wherein I2C communication interface can be used for communicating with FDC2214, and UART communication interface, USB are logical
Letter interface and ethernet mac controller then can be used for communicating with PC machine, to realize summarizing and analyze to gesture data on demand, or
Person's other application extends function.
Common machine learning method has K- neighbour, support vector machines (SVM), neural network, deep learning scheduling algorithm,
Middle K- nearest neighbor algorithm is fairly simple, but when identifying new samples, the learning sample data set before needing to retain, and is calculating
Apart from when, need to calculate the distance between all samples in new samples and learning sample data set, it is therefore desirable to biggish storage
Expense and computing cost.
SVM algorithm is designed for two-value classification problem, and for gesture identification problem of more classifying, it is suitable to need to construct
Multi classifier, computation complexity is high, implements relatively difficult, and training speed is relatively slow, thus is not best
Scheme.
For neural network and deep learning algorithm, a large amount of training data is needed to support, and needs complicated tune ginseng machine
System, this is not equally preferred plan for storage resource and the limited embedded system of computing resource.
Decision Tree algorithms are a kind of typical more classification methods.The algorithm is first handled data, is calculated using concluding
Method generates readable rule and decision tree, is then analyzed using decision new data.Decision tree is substantially by a system
The process that column rule classifies to data.Decision Tree algorithms are high with nicety of grading, generation mode is simple, have to noise data
Preferable stalwartness feature performance benefit, is one of the multi-classification algorithm being most widely used at present.In the training stage, the training number that needs
It is many according to not needing, in cognitive phase, as long as do compare several times as long as can obtain its classification, so regardless of from storage overhead
Above face or computing cost, decision Tree algorithms are all preferably machine learning schemes.
Summary of the invention
The purpose of the present invention is to provide a kind of non-contact gesture identification device and its methods, provide a kind of cost performance
Non-contact gesture identification device high, applied widely, with it is accurate, fast implement gesture identification function in induction region.
The technical solution adopted by the present invention is that:
A kind of non-contact gesture identification device comprising FDC2214 capacitance sensing chip, Spliced type single-side brass plate
And single-chip microcontroller, Spliced type single-side brass plate and gesture to be measured constitute switched capacitances, FDC2214 capacitance sensing chip with can spell
The copper sheet electrical connection of formula single-side brass plate is connect, and detects the switching regulator for obtaining Spliced type single-side brass plate and gesture to be measured composition
Capacitance, FDC2214 capacitance sensing chip output access single-chip microcontroller, and single-chip microcontroller reads FDC2214 capacitance sensing chip and measures
Reflection switched capacitances value frequency data;Single-chip microcontroller has the memory for saving frequency data, and single-chip microcontroller utilizes frequency
Rate data pass through the categorised decision tree of training construction complete difference gesture and using categorised decision tree to the frequency number acquired in real time
According to progress gesture classification and output category result.
Further, single-chip microcontroller is K60 single-chip microcontroller, and K60 single-chip microcontroller passes through the communication interface being arranged thereon and host computer is logical
Letter connection.
Further, communication interface includes UART communication interface, USB communication interface and ethernet interface.
Further, single-side brass plate has four groups of Spliced type metal electrodes, respectively corresponds index finger, middle finger, the third finger
With little finger of toe position, whether and detecting index finger, middle finger, the third finger and little finger of toe and appearing in is four groups of Spliced types above metal electrode
Metal electrode is electrically connected each by a conducting wire with 4 capacitance signal acquisition channels on FDC2214 capacitance sensing chip.
Further, every group of metal electrode is Spliced type, including a long section of lath and at least one short section lath,
Long section lath with short section lath along corresponding finger orientation, successively put by small―gap suture, and long section lath is arranged close to palm position,
Short section lath is arranged far from palm position, and long section lath is connect with short section lath by dismountable conducting bridge, when according to detection
When big palm, more than one short section lath is spliced on the basis of long section lath;When small palm, then disconnecting conducting bridge only makes
With long section lath.
Further, conducting bridge is using conductive metal wire or conductive glue.
Further, the small―gap suture interval 1mm between long section lath and short section lath, the width of long section lath and short section lath
Spending identical is 2cm, and the length of long section lath is 10cm~20cm, and the length of short section lath is 5cm.
Further, corresponding detection middle finger and nameless metal electrode are placed in parallel, its spacing 1cm;Corresponding detection
Angle between the metal electrode of index finger and the metal electrode of middle finger is 10 °, and is divided into 0.5cm between most narrow place;Corresponding inspection
The angle surveyed between the metal electrode of nameless metal electrode and little finger of toe is 10 °, and is divided into 0.5cm between most narrow place.
Further, a kind of non-contact gesture recognition methods comprising following steps:
Step 1, four groups of Spliced type metal electrodes on single-side brass plate are installed according to the size of palm to be detected,
Step 2, single-chip microcontroller acquires several groups gesture data for different gestures respectively and constitutes learning data set;
Step 3, the categorised decision tree of different gestures is obtained based on learning data set training;
Step 3-1 calculates the comentropy of current learning data set D, its calculation formula is:
Wherein pkIndicate kth class gesture sample proportion in whole learning data set D in n class gesture;
Step 3-2, the corresponding attribute of the frequency data of every group of metal electrode, calculates the information gain of all properties,
Calculation formula are as follows:
Wherein, CHx (x=0,1,2,3), CHx indicate xth channel;DvIt is the different values according to CHx to learning data
Collect some subset after D is divided, and has D={ D1, D2..., DV};
Step 3-3, finds out branch foundation of the maximum attribute of information gain as current decision tree node, and the attribute
Child node of each value as present node.
Step 3-4 repeats step 3-1 to step 3-3 to all child nodes, until son node number is all same according to concentrating
Class gesture obtains the categorised decision tree of different gestures in turn;
Step 4, single-chip microcontroller carries out gesture classification and output category to the frequency data acquired in real time using categorised decision tree
As a result.
Further, the specific steps of the step 2 are as follows:
Step 2-1, the several groups frequency data in tetra- channels FDC2214 are formed when single-chip microcontroller reads no gesture state respectively
Primary data;
The gesture data of step 2-2, each gesture that single-chip microcontroller is directed under different application acquire several groups simultaneously respectively
Learning data set is constituted together with primary data.
Further, in step 2-2 learning data set data format according to " gesture data group "+" gesture label " lattice
Formula is stored into the SRAM of single-chip microcontroller.
The invention adopts the above technical scheme, has the advantages that (1) this non-contact gesture identification device uses
FDC2214 capacitance sensor realizes the non contact angle measurement to gesture, opposite to utilize computer vision technique identifying schemes, this hair
It is bright it is at low cost, environmental suitability is strong.(2) gesture identification of the applicable different hand sizes of spliced metal electrode, ideal adaptation
Property is strong.(3) gesture is identified using decision Tree algorithms, accuracy rate is high, and time overhead and storage overhead are all smaller.
Detailed description of the invention
The present invention is described in further details below in conjunction with the drawings and specific embodiments;
Fig. 1 is the detection principle diagram of FDC2214;
Fig. 2 is a kind of structural schematic diagram of non-contact gesture identification device of the present invention.
Specific embodiment
As shown in Fig. 2, the invention discloses a kind of non-contact gesture identification devices comprising FDC2214 capacitance sensing
Chip, Spliced type single-side brass plate and single-chip microcontroller, Spliced type single-side brass plate and gesture to be measured constitute switched capacitances,
FDC2214 capacitance sensing chip is electrically connected with the copper sheet of Spliced type single-side brass plate, and is detected and obtained Spliced type one side applies
The switched capacitances value that copper sheet and gesture to be measured are constituted, FDC2214 capacitance sensing chip output access single-chip microcontroller, and monolithic is machine-readable
The frequency data for the reflection switched capacitances value for taking FDC2214 capacitance sensing chip to measure;Single-chip microcontroller has for saving frequency
The memory of data, single-chip microcontroller pass through the categorised decision tree of training construction complete difference gesture using frequency data and utilize classification
Decision tree carries out gesture classification and output category result to the frequency data acquired in real time.
Specifically, the memory of K60 single-chip microcontroller includes SRAM and FLASH, and frequency data are saved in SRAM by K60 single-chip microcontroller
In, it can also be saved in FLASH as needed, to save after powering off.
Further, single-side brass plate has four groups of Spliced type metal electrodes, respectively corresponds index finger, middle finger, the third finger
With little finger of toe position, whether and detecting index finger, middle finger, the third finger and little finger of toe and appearing in is four groups of metal electrodes above metal electrode
It is electrically connected each by a conducting wire with 4 capacitance signal acquisition channels on FDC2214 capacitance sensing chip.In addition, as one
Feasible embodiment is planted, the gesture acquisition plane in hardware plan can be used a monolith single-side brass plate, draw in the copper clad plate
A channel of conducting wire access FDC2214 out, can allow a FDC2214 4 tunnel gesture identifications of processing simultaneously.
Further, every group of metal electrode includes a long section of lath and at least one short section lath, long section lath with it is short
Section lath along corresponding finger orientation, successively put by small―gap suture, and long section lath is arranged close to palm position, and short section lath is separate
The setting of palm position, long section lath are connect with short section lath by dismountable conducting bridge,
Further, single-chip microcontroller is K60 single-chip microcontroller, and K60 single-chip microcontroller passes through the communication interface being arranged thereon and host computer is logical
Letter connection.Further, communication interface includes UART communication interface, USB communication interface and ethernet interface.Gesture identification
Device can issue real-time or history gesture data PC machine and statistically analyze by these communication interfaces, to increase gesture data
Available time and dubious zone.
Specifically, being adopted for gesture sensing plane since FDC2214 capacitance sensing chip provides 4 capacitance signals
Collect channel, for the hardware resource for making full use of FDC2214, improves gesture motion discrimination, devise four using single-side brass plate
Group metal plate fixes electrode, and whether be orderly used to detection index finger, middle finger, the third finger and little finger of toe and appear in is on metal electrode
Side.
Since the different manpower palms is of different sizes, for example the palm of adult would generally be bigger than the palm of child, the hand of male
Slap it is usually bigger than the palm of women, for the application range for improving gesture identifying device, every group of metal electrode use it is spliced, i.e.,
Different, the copper clad plate item (calling lath in the following text) of same size using two segment length, wherein the length of long lath is according to general palm
Size can be set as 10cm~20cm, and the length of batten ends can be set as 5cm or so, and width is set as 2cm or so.
Firstly, two sections of laths are successively put along certain finger orientation, for long section lath close to palm position, short section lath is remote
From palm position, two sections of lath spacing are as small as possible, but are not directly contacted with, but realize connection on demand by conducting bridge.It is conductive
Bridge can be using conductive metal wire or conductive glue.When detecting larger palm, two sections of laths are connected with conducting bridge,
Collectively as fixed metal electrode plate;When detecting smaller palm, then conducting bridge is disconnected.Under long any situation of section lath all can by with
It arrives, therefore they is respectively drawn to a conducting wire and is connected with four channels of FDC2214 acquisition module.After being opened due to each finger extension simultaneously
It is not parallel, especially little finger of toe and index finger, expanded angle is bigger, therefore for measuring middle finger and nameless fixation metal electricity
Pole group is placed in parallel, and spacing is arranged as needed, such as can be set as half of width of sheet or so;For measuring index finger and little finger of toe
Electrode group then put by 0 °~20 ° of angle by metal electrode group adjacent thereto, but is not contacted, and most narrow place is close to palm one
End, spacing is arranged as needed, such as can be set as the half or so of fixed metal electrode board width.
Every group of metal electrode of gesture sensing plane has been all made of length is respectively 10cm and 5cm, width is 2cm two sections
Single-side brass plate, long batten ends interval 1mm, measurement middle finger and nameless metal electrode group are placed in parallel, and spacing 1cm is surveyed
The angle measured between index finger and middle finger, the third finger and the metal electrode group of little finger of toe is 10 °, is divided into 0.5cm between most narrow place.
Further, a kind of non-contact gesture recognition methods comprising following steps:
Step 1, four groups of metal electrodes on single-side brass plate are installed according to the size of palm to be detected,
Step 2, single-chip microcontroller acquires several groups gesture data for different gestures respectively and constitutes learning data set;
Further, the specific steps of the step 2 are as follows:
Step 2-1, the several groups frequency data in tetra- channels FDC2214 are formed when single-chip microcontroller reads no gesture state respectively
Primary data;
The gesture data of step 2-2, each gesture that single-chip microcontroller is directed under different application acquire several groups simultaneously respectively
Learning data set is constituted together with primary data.Further, in step 2-2 the data format of learning data set according to " gesture
Data group "+" gesture label " format is stored into the SRAM of single-chip microcontroller.
Specifically, in this gesture identifying device scheme, by single-chip microcontroller from the data that FDC2214 is read be frequency number
According to, since the present apparatus need to only distinguish different gestures, the capacitance for knowing gesture sensing plane under different gestures is not needed, because
This, only needs directly to analyze the original frequency data read.
Since the area that different gestures covers above four groups of copper clad plate items is different, thus read from four channels
Frequency data are also different.This four groups of different frequency data can regard the feature of corresponding gesture as.
Gesture data learn the stage, K60 single-chip microcontroller first read original state i.e. without gesture state when FDC2214 tetra- lead to
Road several groups frequency data, the size of data volume, which will be subject to, can accurately distinguish different gestures, for decision Tree algorithms, each hand
Gesture acquires 100 groups of data for learning, and can be obtained 90% or more discrimination.Successively acquisition different application design later
Under gesture data, such as finger-guessing game successively acquires 100 groups " stones ", " scissors ", " cloth " these three gesture datas, for finger-guessing game,
Then successively acquire 100 groups " 1 ", " 2 ", " 3 ", " 4 ", " 5 " this five gesture datas.Primary data and gesture data collection group are closed
Come, constitutes learning data set.Learning data concentrates data format to store according to the format of " gesture data group "+" gesture label "
It comes in the SRAM of K60 single-chip microcontroller, is learnt to give subsequent decision Tree algorithms.Also it can according to need storage to arrive
The FLASH memory of K60 single-chip microcontroller is subject to long-term preservation.The gesture data saved can according to need to be led to by different
Letter interface passes to the end PC.
Step 3, the categorised decision tree of different gestures is obtained based on learning data set training;Learn sample completing gesture data
After the acquisition of this collection, next learn the decision condition of different gestures using decision Tree algorithms.Different gestures, four groups of frequency numbers
According to having differences as, four groups of frequency data can be regarded to four attributes of gesture motion.The data meeting in the same same channel of gesture
There are small differences, can attribute value of the most intermediate value as the same channel of same gesture.Therefore according to the attribute and classification of gesture
Label, so that it may easily learn the attributive character of different gestures by decision Tree algorithms;Specifically processing step is as follows:
Step 3-1 calculates the comentropy of current learning data set D, its calculation formula is:
Wherein pkIndicate kth class gesture sample proportion in whole learning data set D in n class gesture;
Step 3-2, the corresponding attribute of the frequency data of every group of metal electrode, calculates the information gain of all properties,
Calculation formula are as follows:
Wherein, CHx (x=0,1,2,3), CHx indicate xth channel;DvIt is the different values according to CHx to learning data
Collect some subset after D is divided, and has D={ D1, D2..., DV};
Step 3-3, finds out branch foundation of the maximum attribute of information gain as current decision tree node, and the attribute
Child node of each value as present node.
Step 3-4 repeats step 3-1 to step 3-3 to all child nodes, until son node number is all same according to concentrating
Class gesture obtains the categorised decision tree of different gestures in turn;
Due to common gesture quantity and few, for example finger-guessing game gesture is generally three: stone, scissors and cloth, channel data
Also only there are four, therefore, the decision tree that constructs is usually and complicated, and level will not be very deep, is based on above-mentioned categorised decision tree
Gesture identification study needed for the calculating time it is very short, be very suitable in the relatively limited scm application system of resource.In K60
Above-mentioned decision Tree algorithms are realized in single-chip microcontroller, and a categorised decision tree can be obtained.
Step 4, single-chip microcontroller carries out gesture classification and output category to the frequency data acquired in real time using categorised decision tree
As a result.After construction complete gesture categorised decision tree, single-chip microcontroller acquires gesture data in real time and decision tree is transferred to be handled, and passes through
Several times classification results more just can be obtained, and classification results are fed back into user.
Furthermore non-machine learning algorithm can also be used for Gesture Recognition Algorithm as a kind of feasible embodiment,
Since in certain environment, the data of the same gesture usually all relatively, can use each channel data weighted sum, to obtain
It must classify a little.This method can be used for calculating that power is weak, in limited 8 bit single-chip system of memory size.
The invention adopts the above technical scheme, and more communication port based on K60 single-chip microcontroller and sliceable metal electrode are non-to be connect
The hardware plan of touch gesture identifying device.Gesture Recognition Algorithm is divided into study stage and cognitive phase, in study stage benefit
With decision Tree algorithms structural classification decision tree, gesture classification is carried out using the categorised decision tree constructed in cognitive phase.This hair
It is bright to have the advantages that (1) this non-contact gesture identification device connects the non-of gesture using the realization of FDC2214 capacitance sensor
Touch identification, opposite to utilize computer vision technique identifying schemes, the present invention is at low cost, environmental suitability is strong.(2) spliced gold
Belong to the gesture identification of the applicable different hand sizes of electrode, individual adaptability is strong.(3) gesture is known using decision Tree algorithms
Not, accuracy rate is high, and time overhead and storage overhead are all smaller.
Claims (10)
1. a kind of non-contact gesture identification device, it is characterised in that: it includes FDC2214 capacitance sensing chip, Spliced type
Single-side brass plate and single-chip microcontroller, Spliced type single-side brass plate and gesture to be measured constitute switched capacitances, FDC2214 capacitance sensing
Chip is electrically connected with the copper sheet of Spliced type single-side brass plate, and is detected and obtained Spliced type single-side brass plate and gesture structure to be measured
At switched capacitances value, FDC2214 capacitance sensing chip output access single-chip microcontroller, single-chip microcontroller read the sense of FDC2214 capacitor
Survey the frequency data for the reflection switched capacitances value that chip measures;Single-chip microcontroller has the memory for saving frequency data, single
Piece machine passes through the categorised decision tree of training construction complete difference gesture and using categorised decision tree to adopting in real time using frequency data
The frequency data of collection carry out gesture classification and output category result.
2. a kind of non-contact gesture identification device according to claim 1, it is characterised in that: the single-chip microcontroller is K60
Single-chip microcontroller, K60 single-chip microcontroller passes through the communication interface being arranged thereon and host computer communicates to connect.
3. a kind of non-contact gesture identification device according to claim 2, it is characterised in that: the communication interface includes
UART communication interface, USB communication interface and ethernet interface.
4. a kind of non-contact gesture identification device according to claim 1, it is characterised in that: the single-side brass plate tool
There are four groups of Spliced type metal electrodes, four groups of Spliced type metal electrodes respectively correspond index finger, middle finger, the third finger and little finger of toe position
It sets, and whether detect index finger, middle finger, the third finger and little finger of toe and appear in is four groups of Spliced type metal electrodes above metal electrode
It is electrically connected each by a conducting wire with 4 capacitance signal acquisition channels on FDC2214 capacitance sensing chip.
5. a kind of non-contact gesture identification device according to claim 4, it is characterised in that: every group of metal electrode
It is Spliced type, including a long section of lath and at least one short section lath, long section lath is with short section lath along corresponding
Finger orientation successively put by small―gap suture, and long section lath is arranged close to palm position, and short section lath is arranged far from palm position, long section
Lath is connect with short section lath by dismountable conducting bridge, when according to big palm is detected, is spelled on the basis of long section lath
Connect more than one short section lath;When detecting small palm, then conducting bridge is disconnected, only uses long section lath;The conducting bridge is adopted
With conductive metal wire or conductive glue.
6. a kind of non-contact gesture identification device according to claim 5, it is characterised in that: the long section lath with it is short
Small―gap suture interval 1mm between section lath, long section lath and the of same size of short section lath are 2cm, the length of long section lath
For 10cm~20cm, the length of short section lath is 5cm.
7. a kind of non-contact gesture identification device according to claim 5, it is characterised in that: corresponding detection middle finger and nothing
The metal electrode that name refers to is placed in parallel, its spacing 1cm;Between the metal electrode of corresponding detection index finger and the metal electrode of middle finger
Angle be 10 °, and be divided into 0.5cm between most narrow place;The metal electrode of corresponding detection nameless metal electrode and little finger of toe it
Between angle be 10 °, and be divided into 0.5cm between most narrow place.
8. a kind of non-contact gesture recognition methods is applied to a kind of non-contact gesture as claimed in claim 1 to 7 and knows
Other device, it is characterised in that: method the following steps are included:
Step 1, four groups of Spliced type metal electrodes on single-side brass plate are installed according to the size of palm to be detected;
Step 2, single-chip microcontroller acquires several groups gesture data for different gestures respectively and constitutes learning data set;
Step 3, the categorised decision tree of different gestures is obtained based on learning data set training;
Step 3-1 calculates the comentropy of current learning data set D, its calculation formula is:
Wherein pkIndicate kth class gesture sample proportion in whole learning data set D in n class gesture;
Step 3-2, the corresponding attribute of the frequency data of every group of metal electrode, calculates the information gain of all properties, calculates
Formula are as follows:
Wherein, CHx (x=0,1,2,3), CHx indicate xth channel;DvIt is to be drawn according to the different values of CHx to learning data set D
Some subset after point, and have D={ D1, D2..., DV};
Step 3-3 finds out branch foundation of the maximum attribute of information gain as current decision tree node, and the attribute is each
Child node of a value as present node.
Step 3-4 repeats step 3-1 to step 3-3 to all child nodes, until son node number is according to all same class hands of concentration
Gesture obtains the categorised decision tree of different gestures in turn;
Step 4, single-chip microcontroller carries out gesture classification and output category result to the frequency data acquired in real time using categorised decision tree.
9. a kind of non-contact gesture identification device according to claim 8, it is characterised in that: the step 2 it is specific
Step are as follows:
Step 2-1, the several groups frequency data in tetra- channels FDC2214 are formed initially when single-chip microcontroller reads no gesture state respectively
Data;
Step 2-2, single-chip microcontroller be directed to different application under each gesture gesture data acquire respectively several groups and with it is first
Beginning data constitute learning data set together.
10. a kind of non-contact gesture identification device according to claim 9, it is characterised in that: learn number in step 2-2
It is stored according to the format of " gesture data group "+" gesture label " into the SRAM of single-chip microcontroller according to the data format of collection.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910460084.XA CN110275614B (en) | 2019-05-30 | 2019-05-30 | Non-contact gesture recognition device and method thereof |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910460084.XA CN110275614B (en) | 2019-05-30 | 2019-05-30 | Non-contact gesture recognition device and method thereof |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110275614A true CN110275614A (en) | 2019-09-24 |
CN110275614B CN110275614B (en) | 2022-09-30 |
Family
ID=67960378
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910460084.XA Active CN110275614B (en) | 2019-05-30 | 2019-05-30 | Non-contact gesture recognition device and method thereof |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110275614B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112099624A (en) * | 2020-08-25 | 2020-12-18 | 李志斌 | Multimode diamond-shaped frame type capacitive sensing gesture recognition system |
CN113673292A (en) * | 2021-01-14 | 2021-11-19 | 南方科技大学 | Capacitive imaging sensor and gesture form sensing method |
CN114327054A (en) * | 2021-12-21 | 2022-04-12 | 杭州电子科技大学 | Gesture recognition device based on FDC2214 |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN203287847U (en) * | 2013-05-31 | 2013-11-13 | 深圳职业技术学院 | Gesture recognition all-in-one machine based on MGC3130 chip |
KR20150030072A (en) * | 2013-09-11 | 2015-03-19 | 삼성전기주식회사 | Touch sensor to recognize a gesture and controlling method thereof |
KR20170091963A (en) * | 2016-02-02 | 2017-08-10 | 삼성전자주식회사 | Gesture classification apparatus and method using electromyogram signals |
CN107526440A (en) * | 2017-08-28 | 2017-12-29 | 四川长虹电器股份有限公司 | The intelligent electric appliance control method and system of gesture identification based on decision tree classification |
CN109032349A (en) * | 2018-07-10 | 2018-12-18 | 哈尔滨工业大学 | A kind of gesture identification method and system based on millimetre-wave radar |
CN109343694A (en) * | 2018-08-13 | 2019-02-15 | 浙江大学 | A kind of gesture recognition system and method for finger-guessing game finger-guessing game game |
CN109491507A (en) * | 2018-11-14 | 2019-03-19 | 南京邮电大学 | Gesture identifying device based on FDC2214 |
CN109710116A (en) * | 2018-08-23 | 2019-05-03 | 华东师范大学 | A kind of non-contact gesture state recognition system and recognition methods |
CN109799914A (en) * | 2019-03-18 | 2019-05-24 | 大连理工大学 | A kind of gesture identifying device based on FDC2214 |
-
2019
- 2019-05-30 CN CN201910460084.XA patent/CN110275614B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN203287847U (en) * | 2013-05-31 | 2013-11-13 | 深圳职业技术学院 | Gesture recognition all-in-one machine based on MGC3130 chip |
KR20150030072A (en) * | 2013-09-11 | 2015-03-19 | 삼성전기주식회사 | Touch sensor to recognize a gesture and controlling method thereof |
KR20170091963A (en) * | 2016-02-02 | 2017-08-10 | 삼성전자주식회사 | Gesture classification apparatus and method using electromyogram signals |
CN107526440A (en) * | 2017-08-28 | 2017-12-29 | 四川长虹电器股份有限公司 | The intelligent electric appliance control method and system of gesture identification based on decision tree classification |
CN109032349A (en) * | 2018-07-10 | 2018-12-18 | 哈尔滨工业大学 | A kind of gesture identification method and system based on millimetre-wave radar |
CN109343694A (en) * | 2018-08-13 | 2019-02-15 | 浙江大学 | A kind of gesture recognition system and method for finger-guessing game finger-guessing game game |
CN109710116A (en) * | 2018-08-23 | 2019-05-03 | 华东师范大学 | A kind of non-contact gesture state recognition system and recognition methods |
CN109491507A (en) * | 2018-11-14 | 2019-03-19 | 南京邮电大学 | Gesture identifying device based on FDC2214 |
CN109799914A (en) * | 2019-03-18 | 2019-05-24 | 大连理工大学 | A kind of gesture identifying device based on FDC2214 |
Non-Patent Citations (1)
Title |
---|
罗于恒,王洋,刘伟: ""一种非接触式的手势识别装置"", 《科技与创新》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112099624A (en) * | 2020-08-25 | 2020-12-18 | 李志斌 | Multimode diamond-shaped frame type capacitive sensing gesture recognition system |
CN112099624B (en) * | 2020-08-25 | 2022-11-22 | 李志斌 | Multimode diamond-shaped frame type capacitive sensing gesture recognition system |
CN113673292A (en) * | 2021-01-14 | 2021-11-19 | 南方科技大学 | Capacitive imaging sensor and gesture form sensing method |
CN114327054A (en) * | 2021-12-21 | 2022-04-12 | 杭州电子科技大学 | Gesture recognition device based on FDC2214 |
Also Published As
Publication number | Publication date |
---|---|
CN110275614B (en) | 2022-09-30 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110275614A (en) | A kind of non-contact gesture identification device and its method | |
CN105844216B (en) | Detection and matching mechanism for recognizing handwritten letters by WiFi signals | |
CN109492667A (en) | A kind of feature selecting discrimination method for non-intrusive electrical load monitoring | |
CN108182433A (en) | A kind of meter reading recognition methods and system | |
CN106503742A (en) | A kind of visible images insulator recognition methods | |
CN102866179A (en) | Online recognition and inhibition method based on non-target interference smell in electronic nose of artificial intelligent learning machine | |
CN102339389A (en) | Fault detection method for one-class support vector machine based on density parameter optimization | |
CN105929113B (en) | A kind of e-nose signal error adaptive learning method with subspace projection | |
CN109002859B (en) | Sensor array feature selection and array optimization method based on principal component analysis | |
CN110674875A (en) | Pedestrian motion mode identification method based on deep hybrid model | |
CN105117708A (en) | Facial expression recognition method and apparatus | |
Elsts et al. | On-Board Feature Extraction from Acceleration Data for Activity Recognition. | |
CN109359665A (en) | A kind of family's electric load recognition methods and device based on support vector machines | |
CN112633658A (en) | Low-voltage distribution area topological relation identification method based on CNN-LSTM | |
CN109104257A (en) | A kind of wireless signal detection method and device | |
Liao et al. | Recognition of partial discharge patterns | |
CN114359674A (en) | Non-invasive load identification method based on metric learning | |
CN110517730A (en) | A method of thermophilic protein is identified based on machine learning | |
CN109710116B (en) | Non-contact gesture state recognition system and recognition method | |
CN103310205A (en) | Handwritten digit recognition method and device | |
CN110191481A (en) | A kind of method and system using the high-speed decision tree assessing link quality that transfinites | |
CN109343694A (en) | A kind of gesture recognition system and method for finger-guessing game finger-guessing game game | |
Yin et al. | FMI image based rock structure classification using classifier combination | |
CN109191819A (en) | A kind of multiple spot Monitoring And Analysis of The Quality system and method for monitoring and analyzing based on clustering algorithm | |
CN107121482A (en) | A kind of embedded volt-ampere electronic tongues detecting system and method |
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 |