CN108509025A - A kind of crane intelligent Lift-on/Lift-off System based on limb action identification - Google Patents
A kind of crane intelligent Lift-on/Lift-off System based on limb action identification Download PDFInfo
- Publication number
- CN108509025A CN108509025A CN201810078017.7A CN201810078017A CN108509025A CN 108509025 A CN108509025 A CN 108509025A CN 201810078017 A CN201810078017 A CN 201810078017A CN 108509025 A CN108509025 A CN 108509025A
- Authority
- CN
- China
- Prior art keywords
- commanding
- lift
- crane
- limb action
- unit
- 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
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/107—Static hand or arm
- G06V40/113—Recognition of static hand signs
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Human Computer Interaction (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The present invention relates to a kind of crane intelligent Lift-on/Lift-off Systems based on limb action identification, it is characterised in that the man-machine interactive system based on six control units, 1)The big multigroup standard limb action of zoom IP Camera continuous acquisition commanding, 2)Big zoom IP Camera, on crane cockpit, acquires commanding's limb action in real time by tripod head frame, and commanding's gesture Hu image Character eigenvectors are obtained after successively entering contours extract unit and feature extraction unit;3)It is inquired especially by using the sample classification device of SVM machine learning algorithms, returns to most matched limb action sample class, crane hanging component work is completed to which corresponding crane command signal is transmitted to crane controller;It overcomes the shortcomings of above-mentioned existing limb action identification technology, and artificially judge that mode command signal is not friendly enough, it is reference with hoisting commanding signal standard GB/T 5,082 85, can have many advantages, such as that at low cost, strong robustness and real-time are high to commanding's real-time tracing.
Description
Technical field
The present invention relates to a kind of crane intelligent Lift-on/Lift-off Systems based on limb action identification, belong to pattern-recognition and man-machine
Interaction technique field.
Technical background
Commander's gesture that conventional hoist Lift-on/Lift-off System relies on commanding is artificially controlled, with human-computer interaction and figure
As the broad development of processing, combined the intelligent Lift-on/Lift-off System of control for improving lifting efficiency and operator with computer machinery
Comfort has extraordinary application value.
It is had the following deficiencies at this stage using the control method of limb action:
Chinese patent CN105353667A discloses a kind of system and method controlling vehicle by limb action, described
One or several limb action information of human body are pre-deposited data memory module by method as electronic signals, then again
It acquires the limb action information of induction zone intra domain user in real time using sensor device, and limb action information is converted into vehicle
Control instruction.This method is expensive using equipment price, and applied to Lift-on/Lift-off System disadvantage is that operator dress it is complicated
Sensor device is not suitable for real operation environment;
Chinese patent CN102799191A discloses a kind of cloud platform control method and system based on action recognition technology, institute
It states method and obtains the Action Semantic mapping table that preset limb action is converted to standard first, then identification camera is real-time
Operator's limb action information of acquisition, with semanteme reflects color table comparative analysis by the action message of identification and obtains the control of holder and refer to
It enables.This method disadvantage is that semantic-based control method has very strong subjective characteristics;
Chinese patent CN102831380A is disclosed a kind of body action identification method incuded based on depth image and is
System, the method are obtained the deep image information of user and its environment using depth camera, then extract user action limb
Body deep image information is to carry out the identification of action limbs.This method disadvantage is needed using expensive depth camera head apparatus.
Invention content
The purpose of the present invention is to provide it is a kind of based on limb action identification crane intelligent Lift-on/Lift-off System, overcome on
The deficiency of existing limb action identification technology is stated, and artificially judges that mode command signal is not friendly enough, is believed with hoisting commanding
Number standard GB/T 5082-85 is reference, can have at low cost, strong robustness and real-time to commanding's real-time tracing
The advantages that high.
The technical proposal of the invention is realized in this way:A kind of crane intelligent lifting system based on limb action identification
System, it is characterised in that the man-machine interactive system based on six control units, specific implementation include the following steps:
1) the multigroup standard limb action of big zoom IP Camera continuous acquisition commanding, successively enters contours extract list
Member, feature extraction unit and feature training unit train to obtain standard limb maneuver library by SVM machine learning algorithms;
2) it is dynamic to acquire commanding's limbs by tripod head frame on crane cockpit in real time for big zoom IP Camera
Make, commanding's gesture Hu image Character eigenvectors are obtained after successively entering contours extract unit and feature extraction unit;
3) step 2) commanding's gesture Hu image Character eigenvectors feeding characteristic matching unit is inquired most matched
Sample action class is inquired especially by using the sample classification device of SVM machine learning algorithms, and it is dynamic to return to most matched limbs
Make sample class, crane hanging component work is completed to which corresponding crane command signal is transmitted to crane controller;
4) with the movement of hoisting object, commanding also has corresponding movement, and target tracking unit detects commander
Personnel's movement can send holder of the control signal to the big zoom camera of bearer network, and holder is made to track commanding.
Six control units are:
Image acquisition units:For obtaining commanding's image information;
Contours extract unit:For being partitioned into gesture profile information from commanding's image information;
Feature extraction unit:Characteristic information for being partitioned into commander's gesture from commanding;
Feature training unit:For commander's gesture feature information of extraction to be trained;
Characteristic matching unit:Limb action for collection site commanding is acted with trained standard limb
Library carries out match cognization;
Target tracking unit:For after the limb action for identifying floor manager personnel, send out control instruction to
While crane hanging component system, commanding is tracked, ensures commanding always in camera
In monitoring range.
Described image collecting unit is that the big zoom camera of network is that Haikang prestige regards integrated camera (model DS-
2ZCN3007), 30 frames of highest frame per second/second;The camera is fixed on by holder on crane cockpit.
The contours extract unit is to be partitioned into commanding using YCbCr color spaces to command gesture, is used simultaneously
Alpha-beta Kalman filtering excludes the interference of class colour of skin object.
The described limbs gesture to being partitioned into extracts Hu image moment characteristic informations.
The feature training unit is that each standard command and control gesture corresponds to one using SVM machine learning algorithms
Sample classification device, the sample classification implement body puies forward the standard command and control gesture by using SVM machine learning algorithms
The Hu image Character eigenvector classification learnings taken obtain.
The characteristic matching unit is that real-time collection site commanding commands gesture, is looked into standard limb maneuver library
Most matched sample action class is ask out, is inquired especially by using the sample classification device of SVM machine learning algorithms, is returned most
Matched sample action class, the sample classification implement body is by using SVM machine learning algorithms to limb action Hu image moments
Feature vector classification learning obtains.
The target tracking unit moves for commanding with the movement of hoisting object, believes in commanding's image
In breath, in order to ensure that commanding at the center of image, chases after the target of commanding using the completion of CamShift algorithms always
Track, i.e., the pixel coordinate offset that commanding is deviateed to picture centre are converted to the angle of horizontal stage electric machine movement.
Good effect of the present invention be compared with conventional hoist Lift-on/Lift-off System rely on commanding commander's gesture artificially controlled
System is combined machinery with computer the intelligent Lift-on/Lift-off System of control to have very for improving lifting efficiency and operator's comfort
Good application value;Present system can have the advantages that robustness is high to commanding's real-time tracing simultaneously.
Description of the drawings
Fig. 1 crane hanging component system general illustrations.
Fig. 2 holders, big zoom IP Camera and cockpit partial enlarged view.
Fig. 3 overview flow charts of the present invention.
Fig. 4 contours extract cell operation schematic diagrams.
Fig. 5 target tracking cell operation schematic diagrams.
Specific implementation mode
The present invention is described in further details with reference to the accompanying drawings and examples.A kind of rising based on limb action identification
Heavy-duty machine intelligence Lift-on/Lift-off System, including image acquisition units, contours extract unit, feature extraction unit, feature training unit, feature
Matching unit and target tracking unit;It is characterized in that it specifically includes following steps:
It first passes around SVM machine learning algorithms to train to obtain standard limb maneuver library, wherein standard limb action is with state
Family standard GB5082-85 is reference;Then it acquires commanding's limb action in real time by image acquisition units, successively enters profile
Commanding's gesture Hu image Character eigenvectors are obtained after extraction unit and feature extraction unit, which is sent into characteristic matching
Unit inquires most matched limb action sample class, complete to which corresponding crane command signal is transmitted to crane controller
It works at crane hanging component, while target tracking unit completes the tracking to commanding.
Wherein image acquisition units:For obtaining commanding's image information;
Contours extract unit:For being partitioned into gesture profile information from commanding's image information;
Feature extraction unit:Gesture profile information for being partitioned into from commanding extracts image moment characteristic information, carries
The characteristic information taken is trained to be delivered to feature training unit;
Feature training unit:For the image moment characteristic information of extraction to be trained, standard limb maneuver library is obtained;
Characteristic matching unit:For collection site commanding limb action and trained standard limb maneuver library into
Row match cognization;
Target tracking unit:For after the limb action for identifying floor manager personnel, sending out control instruction to lifting
While machine Lift-on/Lift-off System, commanding is tracked, ensures commanding always within the scope of camera head monitor.
It is that Haikang prestige regards integrated camera (model DS- that described image collecting unit, which is the big zoom camera of network,
2ZCN3007), 30 frames of highest frame per second/second;The camera is fixed on by holder on crane cockpit.
The contours extract unit is partitioned into commanding using YCbCr color spaces and commands gesture, while using alpha-beta
Kalman filtering excludes the interference of class colour of skin object.
The feature extraction unit extracts Hu image moment characteristic informations to the limbs gesture being partitioned into.
The feature training unit, each standard command and control gesture correspond to one using SVM machine learning algorithms
Sample classification device, sample classification implement body are extracted to the standard command and control gesture by using SVM machine learning algorithms
Hu image Character eigenvector classification learnings obtain.
The characteristic matching unit, real-time collection site commanding command gesture, are inquired in standard limb maneuver library
Go out most matched sample action class, inquired especially by using the sample classification devices of SVM machine learning algorithms, is returned most
The sample action class matched, the sample classification implement body are special to limb action Hu image moments by using SVM machine learning algorithms
The vectorial classification learning of sign obtains.
The target tracking unit completes the target tracking to commanding using CamShift algorithms, ensures commander people
Member is always in the center of monitoring range.
As shown in Figs. 1-2, commanding 1 is towards big zoom IP Camera 4, commanding by adjust suitable position with
It avoids interfering with hoisting object 3.Big zoom IP Camera locks the position of commanding by Face datection, wherein big zoom
IP Camera is fixed on by sliding slot on cradle head of two degrees of freedom 5.Holder, which has, to be rotated horizontally and upper and lower pitching movement two
Degree of freedom can efficiently accomplish the locking to commanding, and wherein holder is fixed on by holder pedestal 6 on cockpit 7.In order to
The position of big zoom IP Camera, holder and cockpit is clearly expressed referring to Fig. 2 holders, big zoom IP Camera and driving
Cabin partial enlarged view.
For overview flow chart of the present invention referring to Fig. 3, flow chart is divided into two big modules:Training module and operational module.Training mould
The multigroup standard limb action of the big zoom camera continuous acquisition commanding of network, successively enters contours extract unit, spy in block
Sign extraction unit and feature training unit train to obtain standard limb maneuver library by SVM machine learning algorithms;In operational module
Big zoom IP Camera (by tripod head frame on crane cockpit) acquires commanding's limb action in real time, successively enters
Commanding's gesture Hu image Character eigenvectors are obtained after contours extract unit and feature extraction unit, which is sent into feature
Matching unit inquires most matched sample action class, is carried out especially by using the sample classification device of SVM machine learning algorithms
Inquiry returns to most matched sample action class, is completed to which corresponding crane command signal is transmitted to crane controller
Heavy-duty machine lifts work.With the movement of hoisting object, commanding also has corresponding movement, and target tracking unit detects finger
The holder that personnel's movement can send control signal to the big zoom camera of bearer network is waved, holder is made to track commanding.
In the above method, system specifically includes following six unit:
1) image acquisition units:For obtaining commanding's image information;
2) contours extract unit:For being partitioned into gesture profile information from commanding's image information;
3) feature extraction unit:Gesture profile information for being partitioned into from commanding extracts image moment characteristic information;
4) feature training unit:For the image moment characteristic information of extraction to be trained, standard limb maneuver library is obtained;
5) characteristic matching unit:Limb action for collection site commanding and trained standard limb maneuver library
Carry out match cognization;
6) target tracking unit:For after the limb action for identifying floor manager personnel, send out control instruction to
While heavy-duty machine Lift-on/Lift-off System, commanding is tracked, ensures commanding always within the scope of camera head monitor.
The principle of contours extract unit mentioned above and target tracking unit is further explained in detail below:
Described its operating diagram of contours extract unit has brightness and color referring to Fig. 4, by the YCbCr color space colours of skin
Mutually independent characteristic is spent, the present invention analyzes limbs Hand Gesture Segmentation using YCbCr color spaces.Acquisition is commanded first
Personnel's image information is transformed into YCbCr color spaces with following formula by rgb space:
Wherein Y refers to luminance component image, and Cb indicates that image blue component, Cr indicate image red component.To skin color range
The spaces CbCr be split, consult pertinent literature range Cr=[133,173], Cb=[77,127].To the hand being partitioned into
Gesture excludes the interference of class colour of skin object using alpha-beta Kalman filtering, and it is adjacent to be connected to then to carry out morphology closed operation processing again
Agglomerate and filling minuscule hole, obtain the full clearly gesture of profile.
Described its calculation amount of alpha-beta Kalman filtering is smaller, and principle is that its state equation is fixed as two compared with Kalman filtering
Dimension.If being x in the state at k-1 momentk-1, it is assumed that sampling interval T then predicts the state estimation at k momentFor:
WhereinIt is respectively t=k-1 moment speed of moving body and position with x (k-1),WithPoint
It is notWith the estimated value of x (k-1).The then velocity estimation value at K momentFor
Two formulas are alpha-beta state equation above.It is not anti-to set m displacement measurement of k moment as xm(k), exist due to measuring
Strong noise, at this time we can obtain k moment speed and the estimated state value of displacement:
Wherein α (k-1) and β (k-1) is the filtering gain at k-1 moment, and above formula shows speed and state estimation when t=kWithIt is by state errorIt corrects.A α-βfilter is formed by following four formulas:
Wherein,It representsSquare of mean value, σm 2Represent variance.
Referring to Fig. 5, commanding moves described its operating diagram of target tracking unit with the movement of hoisting object,
In commanding's image information, in order to ensure commanding always at the center of image, the present invention utilizes CamShift algorithms
Target tracking is carried out to commanding, i.e., the pixel coordinate offset that commanding is deviateed to picture centre is converted to horizontal stage electric machine
The angle of movement.
The CamShift algorithms can adjust search window in real time during tracking with the variation of target sizes
Size still finds optimal iteration result for each frame in video sequence using MeanShift.It is calculated in MeanShift
The zeroth order square M that commanding's barycenter uses window is found in method00With first moment M10And M01:
Wherein (x, y) is the coordinate of image, and I (x, y) represents the gray value of (x, y) point.Search window can be obtained by above formula
Center-of-mass coordinate is (xc,yc)=(M10/M00,M01/M00)。
Claims (7)
1. a kind of crane intelligent Lift-on/Lift-off System based on limb action identification, it is characterised in that the people based on six control units
Machine interactive system, specific implementation include the following steps:
1) the multigroup standard limb action of big zoom IP Camera continuous acquisition commanding, successively enter contours extract unit,
Feature extraction unit and feature training unit train to obtain standard limb maneuver library by SVM machine learning algorithms;
2) big zoom IP Camera is by tripod head frame on crane cockpit, and acquisition commanding's limb action, first in real time
Commanding's gesture Hu image Character eigenvectors are obtained after entering contours extract unit and feature extraction unit afterwards;
3) step 2) commanding's gesture Hu image Character eigenvectors feeding characteristic matching unit is inquired into most matched action
Sample class is inquired especially by using the sample classification device of SVM machine learning algorithms, returns to most matched limb action sample
This class completes crane hanging component work to which corresponding crane command signal is transmitted to crane controller;
4) with the movement of hoisting object, commanding also has corresponding movement, and target tracking unit detects commanding
Movement can send holder of the control signal to the big zoom camera of bearer network, and holder is made to track commanding.
2. according to a kind of crane intelligent Lift-on/Lift-off System based on limb action identification described in claim 1, feature exists
It is in six control units:
Image acquisition units:For obtaining commanding's image information;
Contours extract unit:For being partitioned into gesture profile information from commanding's image information;
Feature extraction unit:Characteristic information for being partitioned into commander's gesture from commanding;
Feature training unit:For commander's gesture feature information of extraction to be trained;
Characteristic matching unit:Limb action for collection site commanding and the progress of trained standard limb maneuver library
With identification;
Target tracking unit:For after the limb action for identifying floor manager personnel, sending out control instruction and being hung to crane
While dress system, commanding is tracked, ensures commanding always within the scope of camera head monitor.
3. according to a kind of crane intelligent Lift-on/Lift-off System based on limb action identification described in claim 1, feature exists
In described image collecting unit be the big zoom camera of network be Haikang prestige regard integrated camera (model DS-2ZCN3007),
30 frames of highest frame per second/second;The camera is fixed on by holder on crane cockpit.
4. according to a kind of crane intelligent Lift-on/Lift-off System based on limb action identification described in claim 1, feature exists
It is to be partitioned into commanding using YCbCr color spaces to command gesture, while using alpha-beta karr in the contours extract unit
Graceful filtering excludes the interference of class colour of skin object.
5. according to a kind of crane intelligent Lift-on/Lift-off System based on limb action identification described in claim 1, feature exists
In the limbs gesture to being partitioned into, Hu image moment characteristic informations are extracted.
6. according to a kind of crane intelligent Lift-on/Lift-off System based on limb action identification described in claim 1, feature exists
It is that each standard command and control gesture corresponds to a sample using SVM machine learning algorithms in the feature training unit
Grader, the Hu that the sample classification implement body extracts the standard command and control gesture by using SVM machine learning algorithms
Image Character eigenvector classification learning obtains.
The characteristic matching unit is that real-time collection site commanding commands gesture, is inquired in standard limb maneuver library
Most matched sample action class is inquired especially by using the sample classification device of SVM machine learning algorithms, and return most matches
Sample action class, the sample classification implement body is by using SVM machine learning algorithms to limb action Hu image moment characteristics
Vectorial classification learning obtains.
7. according to a kind of crane intelligent Lift-on/Lift-off System based on limb action identification described in claim 1, feature exists
It is moved with the movement of hoisting object for commanding in the target tracking unit, in commanding's image information,
In order to ensure that commanding always at the center of image, completes the target tracking to commanding, i.e., using CamShift algorithms
The pixel coordinate offset that commanding is deviateed to picture centre is converted to the angle of horizontal stage electric machine movement.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810078017.7A CN108509025A (en) | 2018-01-26 | 2018-01-26 | A kind of crane intelligent Lift-on/Lift-off System based on limb action identification |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810078017.7A CN108509025A (en) | 2018-01-26 | 2018-01-26 | A kind of crane intelligent Lift-on/Lift-off System based on limb action identification |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108509025A true CN108509025A (en) | 2018-09-07 |
Family
ID=63375564
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810078017.7A Pending CN108509025A (en) | 2018-01-26 | 2018-01-26 | A kind of crane intelligent Lift-on/Lift-off System based on limb action identification |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108509025A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110061755A (en) * | 2019-04-30 | 2019-07-26 | 徐州重型机械有限公司 | Operation householder method and system, wearable device and engineering truck |
CN110654980A (en) * | 2019-09-27 | 2020-01-07 | 三一海洋重工有限公司 | Crane control system and operation method thereof |
CN111240486A (en) * | 2020-02-17 | 2020-06-05 | 河北冀联人力资源服务集团有限公司 | Data processing method and system based on edge calculation |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102339379A (en) * | 2011-04-28 | 2012-02-01 | 重庆邮电大学 | Gesture recognition method and gesture recognition control-based intelligent wheelchair man-machine system |
CN102930270A (en) * | 2012-09-19 | 2013-02-13 | 东莞中山大学研究院 | Method and system for identifying hands based on complexion detection and background elimination |
CN103530892A (en) * | 2013-10-21 | 2014-01-22 | 清华大学深圳研究生院 | Kinect sensor based two-hand tracking method and device |
CN103686107A (en) * | 2013-12-13 | 2014-03-26 | 华为技术有限公司 | Processing method and device based on projected image |
-
2018
- 2018-01-26 CN CN201810078017.7A patent/CN108509025A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102339379A (en) * | 2011-04-28 | 2012-02-01 | 重庆邮电大学 | Gesture recognition method and gesture recognition control-based intelligent wheelchair man-machine system |
CN102930270A (en) * | 2012-09-19 | 2013-02-13 | 东莞中山大学研究院 | Method and system for identifying hands based on complexion detection and background elimination |
CN103530892A (en) * | 2013-10-21 | 2014-01-22 | 清华大学深圳研究生院 | Kinect sensor based two-hand tracking method and device |
CN103686107A (en) * | 2013-12-13 | 2014-03-26 | 华为技术有限公司 | Processing method and device based on projected image |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110061755A (en) * | 2019-04-30 | 2019-07-26 | 徐州重型机械有限公司 | Operation householder method and system, wearable device and engineering truck |
CN110061755B (en) * | 2019-04-30 | 2021-04-16 | 徐州重型机械有限公司 | Operation assisting method and system, wearable device and engineering vehicle |
CN110654980A (en) * | 2019-09-27 | 2020-01-07 | 三一海洋重工有限公司 | Crane control system and operation method thereof |
CN110654980B (en) * | 2019-09-27 | 2021-12-17 | 三一海洋重工有限公司 | Crane control system and operation method thereof |
CN111240486A (en) * | 2020-02-17 | 2020-06-05 | 河北冀联人力资源服务集团有限公司 | Data processing method and system based on edge calculation |
CN111240486B (en) * | 2020-02-17 | 2021-07-02 | 河北冀联人力资源服务集团有限公司 | Data processing method and system based on edge calculation |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106598226B (en) | A kind of unmanned plane man-machine interaction method based on binocular vision and deep learning | |
CN105739702B (en) | Multi-pose finger tip tracking for natural human-computer interaction | |
CN107660039B (en) | A kind of lamp control system of identification dynamic gesture | |
CN109949361A (en) | A kind of rotor wing unmanned aerial vehicle Attitude estimation method based on monocular vision positioning | |
CN104123549B (en) | Eye positioning method for real-time monitoring of fatigue driving | |
CN108942923A (en) | A kind of mechanical arm crawl control method | |
CN108509025A (en) | A kind of crane intelligent Lift-on/Lift-off System based on limb action identification | |
Tara et al. | Hand segmentation from depth image using anthropometric approach in natural interface development | |
CN108960067A (en) | Real-time train driver motion recognition system and method based on deep learning | |
CN104200200B (en) | Fusion depth information and half-tone information realize the system and method for Gait Recognition | |
CN105868574B (en) | A kind of optimization method of camera track human faces and wisdom health monitor system based on video | |
Hongo et al. | Focus of attention for face and hand gesture recognition using multiple cameras | |
WO2016033279A1 (en) | Method and system of temporal segmentation for gesture analysis | |
KR20150021526A (en) | Self learning face recognition using depth based tracking for database generation and update | |
CN106503651B (en) | A kind of extracting method and system of images of gestures | |
CN106874884A (en) | Human body recognition methods again based on position segmentation | |
CN108446586B (en) | Method for detecting specific action of train driver | |
CN109558825A (en) | A kind of pupil center's localization method based on digital video image processing | |
CN104281839A (en) | Body posture identification method and device | |
US10803604B1 (en) | Layered motion representation and extraction in monocular still camera videos | |
CN106814853A (en) | A kind of eye control tracking based on machine learning | |
Guo | Research of hand positioning and gesture recognition based on binocular vision | |
CN109726681A (en) | It is a kind of that location algorithm is identified based on the blind way of machine learning identification and image segmentation | |
CN108681403A (en) | A kind of trolley control method using eye tracking | |
CN110021029A (en) | A kind of real-time dynamic registration method and storage medium suitable for RGBD-SLAM |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180907 |
|
RJ01 | Rejection of invention patent application after publication |