CN109086731A - It is a kind of for carrying out the robot and storage medium of behavior monitoring - Google Patents
It is a kind of for carrying out the robot and storage medium of behavior monitoring Download PDFInfo
- Publication number
- CN109086731A CN109086731A CN201810927706.0A CN201810927706A CN109086731A CN 109086731 A CN109086731 A CN 109086731A CN 201810927706 A CN201810927706 A CN 201810927706A CN 109086731 A CN109086731 A CN 109086731A
- Authority
- CN
- China
- Prior art keywords
- target person
- image
- behavior
- face
- determination
- 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
- 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
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
-
- 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Psychiatry (AREA)
- Social Psychology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Signal Processing (AREA)
- Collating Specific Patterns (AREA)
- Image Analysis (AREA)
Abstract
The present invention provides a kind of for carrying out the robot and storage medium of behavior monitoring, wherein, the robot includes memory, processor and image collecting device, memory is for storing executable program code and data, the executable program code that processor is used to that memory to be called to store executes following steps: control image collecting device carries out Image Acquisition to target person, to obtain the image of target person, the image of target person has included at least the posture of target person and the face of target person;It is identified based on image of the posture of target person to target person, so that it is determined that the behavior of target person;Face based on target person identifies the image of target person, so that it is determined that the identity of target person;If the behavior of target person is non-permitted behavior, the non-permitted behavior record is in the corresponding data base entries of identity of target person.Using the present invention, can the behavior automatically to on-site employee be monitored, improve monitoring efficiency.
Description
Technical field
The present invention relates to artificial intelligence, and in particular to a kind of for carrying out the robot and storage medium of behavior monitoring.
Background technique
As the application demand of robot is continuously increased, artificial intelligence the relevant technologies are constantly progressive, the growth of hardware performance,
Service robot starts to move towards factory from laboratory in recent years, and develops from simple function to multifunctional personal robot.It mentions
To robot, a word often referred to recently is artificial intelligence.Artificial intelligence is the intelligence realized with computer similar to people
One subject of energy behavior.Robot itself is one of ultimate application target for artificial intelligence.
Traditional artificial intelligence is as a subject, the Dartmouth meeting originating from the 1950s, passes through later
It rises and fall sharply and quickly several times, achievement abundant is had accumulated in basic theory and method.From the Symbolic Computation System of early stage, it is to expert
System, then to the machine learning that the nineties grows up, big data analysis can be the scope of artificial intelligence.In image, language
The fields such as sound, search, data mining, social computing, and derived some relevant application studies.Wherein contacted with robot
It include closely more computer vision, voice and natural language processing, there are also intelligent bodies (Agent) etc..
It can consider following perception, cognition according to the progress of previous robot field and to the preliminary analysis of application
Technology will realize application.
1, three-dimensional navigation location technology.Regardless of robot, as long as removable, that is, need in family or other environment
Carry out navigator fix.Wherein SLAM (Simultaneous Localization and Ming) technology can carry out simultaneously positioning and
Figure is built, there are many technological accumulation in terms of academic research.But for real system, due to real-time low cost (such as nothing
Method uses more expensive radar equipment) requirement and home environment dynamic change (putting for article), thus it is fixed to navigation
Position technology proposes requirements at the higher level, still needs to further research and develop.
2, visual perception technology.It wherein include recognition of face, gesture identification, object identification skill related to Emotion identification etc.
Art.Visual perception technology is a very important technology of robot and people's interaction.
3, language interaction technique.It wherein include speech recognition, speech production, natural language understanding and Intelligent dialogue system
Deng.
4, character recognition technology.There are many text informations in life, such as the label information of books and newspapers and object, this also requires machine
Device people can carry out Text region by camera.With after traditional scanning identify text compared with, can by camera come
Carry out the identification of text.
5, cognitive techniques.Robot needs to be done step-by-step the cognitive functions such as planning, reasoning, memory, study and prediction, thus
Become more intelligent.
In terms of current present Research, the key technology that service robot faces has rapid progress, but there are also quite
More problems will solve.
Summary of the invention
The embodiment of the invention provides a kind of behavior monitoring method and robots, storage medium, can be automatically to plant area
The behavior of interior employee is monitored, and improves monitoring efficiency.
The purpose of the embodiment of the present invention is that be achieved through the following technical solutions:
A kind of behavior monitoring method, comprising:
Image Acquisition is carried out to target person, to obtain the image of the target person, the image of the target person is extremely
It less include the posture of the target person and the face of the target person;
Posture based on the target person identifies the image of the target person, with the determination target person
Behavior;
Face based on the target person identifies the image of the target person, with the determination target person
Identity;
If the behavior of the target person is non-permitted behavior, the non-permitted behavior record is the target person
In the corresponding data base entries of identity of member.
Optionally, the posture based on the target person identifies the image of the target person, with determination
The behavior of the target person includes:
Posture based on the target person extracts the behavioural characteristic of the target person;
Based on the behavioural characteristic, the behavior of the target person is identified using pre-set algorithm.
Optionally, the face based on the target person identifies the image of the target person, with determination
The identity of the target person includes:
Face based on the target person extracts the facial characteristics of the target person;
It is searched from database according to the facial characteristics, with the identity of the determination target person.
Optionally, described that Image Acquisition is carried out to target person, include: to obtain the image of the target person
Image Acquisition is carried out to the target person from multiple angles, to obtain the target person with different angle
Multiple images;
The face based on the target person identifies the image of the target person, with the determination target
The identity of personnel includes:
Multiple images based on the target person carry out three-dimensional modeling to the face of the target person, described in obtaining
The threedimensional model of the face of target person;
The facial characteristics of the target person is determined based on the threedimensional model of the face of the target person;
It is searched from database according to the facial characteristics, with the identity of the determination target person.
Optionally, the quantity of the target person is at least two, the behavior of the target person of the determination be it is described extremely
The combination behavior of few two target persons.
It is a kind of for carrying out the robot of behavior monitoring, comprising:
Image acquisition units, it is described to obtain the image of the target person for carrying out Image Acquisition to target person
The image of target person has included at least the posture of the target person and the face of the target person;
Behavior determination unit identifies for image of the posture based on the target person to the target person,
With the behavior of the determination target person;
Identity determination unit, for based on the target person face the image of the target person is identified,
With the identity of the determination target person;
Processing unit, when for determining that the behavior of the target person is non-permitted behavior in the behavior determination unit,
The non-permitted behavior record is then to the corresponding data of identity of the determining target person of the identity determination unit
In the entry of library.
Optionally, the behavior determination unit is specifically used for:
Posture based on the target person extracts the behavioural characteristic of the target person;
Based on the behavioural characteristic, the behavior of the target person is identified using pre-set algorithm.
Optionally, the identity determination unit is specifically used for:
Face based on the target person extracts the facial characteristics of the target person;
It is searched from database according to the facial characteristics, with the identity of the determination target person.
Optionally, described image acquisition unit is specifically used for:
Image Acquisition is carried out to the target person from multiple angles, to obtain the target person with different angle
Multiple images;
The identity determination unit is specifically used for:
Multiple images based on the target person carry out three-dimensional modeling to the face of the target person, described in obtaining
The threedimensional model of the face of target person;
The facial characteristics of the target person is determined based on the threedimensional model of the face of the target person;
It is searched from database according to the facial characteristics, with the identity of the determination target person.
Optionally, the quantity of the target person is two or more, and the behavior of the target person of the determination is described
The combination behavior of at least two target persons.
A kind of computer readable storage medium, the storage medium are stored with computer program, the computer program quilt
Processor performs the steps of when executing
Image Acquisition is carried out to target person, to obtain the image of the target person, the image of the target person is extremely
It less include the posture of the target person and the face of the target person;
Posture based on the target person identifies the image of the target person, with the determination target person
Behavior;
Face based on the target person identifies the image of the target person, with the determination target person
Identity;
If the behavior of the target person is non-permitted behavior, the non-permitted behavior record is the target person
In the corresponding data base entries of identity of member.
Optionally, posture of the processor based on the target person identifies the image of the target person,
Include: in a manner of the behavior of the determination target person
Posture based on the target person extracts the behavioural characteristic of the target person;
Based on the behavioural characteristic, using behavior of the pre-set algorithm to the target person in the storage medium
It is identified.
Optionally, the processor identifies the image of the target person based on the face of the target person,
Include: in a manner of the identity of the determination target person
Face based on the target person extracts the facial characteristics of the target person;
It is searched from database according to the facial characteristics, with the identity of the determination target person.
Optionally, the processor carries out Image Acquisition to target person, to obtain the side of the image of the target person
Formula includes:
Image Acquisition is carried out to the target person from multiple angles, to obtain the target person with different angle
Multiple images;
Wherein, the processor identifies the image of the target person based on the face of the target person, with
The mode for determining the identity of the target person includes:
Multiple images based on the target person carry out three-dimensional modeling to the face of the target person, described in obtaining
The threedimensional model of the face of target person;
The facial characteristics of the target person is determined based on the threedimensional model of the face of the target person;
It is searched from database according to the facial characteristics, with the identity of the determination target person.
Optionally, the quantity of the target person is at least two, the behavior of the target person of the determination be it is described extremely
The combination behavior of few two target persons.
It is a kind of for carrying out the robot of behavior monitoring, comprising: memory, processor and image collecting device, wherein institute
Memory is stated for storing executable program code and data, described image acquisition device is for acquiring image, the processor
For calling the executable program code of the memory storage, execution following steps:
It controls described image acquisition device and Image Acquisition is carried out to target person, to obtain the image of the target person,
The image of the target person has included at least the posture of the target person and the face of the target person;
Posture based on the target person identifies the image of the target person, with the determination target person
Behavior;
Face based on the target person identifies the image of the target person, with the determination target person
Identity;
When the behavior for determining the target person is non-permitted behavior, then the non-permitted behavior record is described
In the corresponding data base entries of identity of the target person of memory storage.
Optionally, posture of the processor based on the target person identifies the image of the target person,
Include: in a manner of the behavior of the determination target person
Posture based on the target person extracts the behavioural characteristic of the target person;
Based on the behavioural characteristic, using pre-set algorithm in the memory to the behavior of the target person into
Row identification.
Optionally, the processor identifies the image of the target person based on the face of the target person,
Include: in a manner of the identity of the determination target person
Face based on the target person extracts the facial characteristics of the target person;
It is searched from the database that the memory stores according to the facial characteristics, with the determination target person
Identity.
Optionally, the processor control described image acquisition device carries out Image Acquisition to target person, to obtain
The mode for stating the image of target person includes:
It controls described image acquisition device and Image Acquisition is carried out to the target person from multiple angles, have not to obtain
With the multiple images of the target person of angle;
Wherein, the processor identifies the image of the target person based on the face of the target person, with
The mode for determining the identity of the target person includes:
Multiple images based on the target person carry out three-dimensional modeling to the face of the target person, described in obtaining
The threedimensional model of the face of target person;
The facial characteristics of the target person is determined based on the threedimensional model of the face of the target person;
It is searched from the database that the memory stores according to the facial characteristics, with the determination target person
Identity.
Optionally, the quantity of the target person is two or more, and the behavior of the target person of the determination is described
The combination behavior of at least two target persons.
From the above it can be seen that using provided in an embodiment of the present invention for carrying out the robot of behavior monitoring, it can be to target person
Member carries out Image Acquisition, and to obtain the image of the target person, the image of the target person has included at least the target
The face of the posture of personnel and the target person;Posture based on the target person carries out the image of the target person
Identification, so that it is determined that the behavior of the target person;Face based on the target person to the image of the target person into
Row identification, so that it is determined that the identity of the target person;If the behavior of the target person is non-permitted behavior, will be described non-
Behavior record is allowed to be in the corresponding data base entries of the identity of the target person, to improve the efficiency of behavior monitoring;
Further, it when the behavior of the user is hazardous act, can alarm.
Detailed description of the invention
In order to illustrate the technical solution of the embodiments of the present invention more clearly, required use in being described below to embodiment
Attached drawing be briefly described, it should be apparent that, drawings in the following description are only some embodiments of the invention, for this
For the those of ordinary skill of field, without any creative labor, it can also be obtained according to these attached drawings other
Attached drawing.
Fig. 1 is the flow chart of behavior monitoring method provided by one embodiment of the present invention;
Fig. 2 is provided by one embodiment of the present invention for carrying out the structure chart of the robot of behavior monitoring;
Fig. 3 is the structure chart for the robot for carrying out behavior monitoring that another embodiment of the present invention provides.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
Behavior monitoring method provided in an embodiment of the present invention is first introduced, Fig. 1 describes provided by one embodiment of the present invention
The process of behavior monitoring method.Wherein, this method can be applied to robot.As shown in Figure 1, this method may include:
101, Image Acquisition is carried out to target person, to obtain the image of the target person, the figure of the target person
As having included at least the posture of the target person and the face of the target person.
The target person refers to the personnel in workplace, such as office, meeting room, workshop etc..
In one embodiment, when carrying out Image Acquisition to target person, an image can be only acquired, to be based on
The image recognition technology of 2D carries out face recognition.It can complete to identify than faster in this way, to improve identification
Efficiency.
In another embodiment, in order to improve the accuracy of detection, the image recognition technology of 3D can be used, this
When, Image Acquisition is carried out to the target person from multiple angles, to obtain the more of the target person with different angle
A image.
102, the face based on the target person identifies the image of the target person, with the determination target
The identity of personnel.
Wherein, in one embodiment, the image recognition technology of 2D is used, at this point it is possible to described in acquisition
The image of personnel carries out the extraction of facial characteristics, and the facial characteristics includes but is not limited to: the shape of face, between face
Distance etc..Then the face based on the target person identifies the image of the target person, so that it is determined that
The identity of the target person may include:
Face based on the target person extracts the facial characteristics of the target person;
It is searched from database according to the facial characteristics, so that it is determined that the identity of the target person.
In another embodiment, the image recognition technology of 3D is used, at this point, firstly the need of the personnel are based on
Multiple images to the personnel face carry out three-dimensional modeling, thus obtain the personnel face threedimensional model;Then
The facial characteristics of the personnel is determined based on the threedimensional model of the face of the personnel.At this point, the facial characteristics of the personnel removes
It may include the shape of face, it can also include the height of nose that the distance between face etc. are outer, the depth of eye socket, under
Bar convex-concave degree etc..It is described that Image Acquisition is carried out to target person, include: to obtain the image of the target person
Image Acquisition is carried out to the target person from multiple angles, to obtain the target person with different angle
Multiple images;
The face based on the target person identifies the image of the target person, with the determination target
The identity of personnel includes:
Multiple images based on the target person carry out three-dimensional modeling to the face of the target person, described in obtaining
The threedimensional model of the face of target person;
The facial characteristics of the target person is determined based on the threedimensional model of the face of the target person;
It is searched from database according to the facial characteristics, with the identity of the determination target person.
Wherein, existing image recognition technology can be used in image recognition technology, and the embodiment of the present invention also can be used and mention
The image recognition technology of confession.Image recognition technology provided in an embodiment of the present invention be realized based on neural network model, such as
Can be convolutional neural networks (CNN:Convolutional Neural Network) model or Recognition with Recurrent Neural Network (RNN:
Recurrent Neural Networks) model, or it is also possible to deep neural network (DNN:Deep Neural
Networks) model.
103, the posture based on the target person identifies the image of the target person, with the determination target
The behavior of personnel.If the behavior of the target person is non-permitted behavior, 104 are entered step;If the behavior of the target person
To allow behavior, 105 are entered step.
Wherein, the posture based on the target person identifies the image of the target person, so that it is determined that
The behavior of the target person may include:
Posture based on the target person extracts the behavioural characteristic of the target person;
Based on the behavioural characteristic, the behavior of the target person is identified using pre-set algorithm.
Wherein, such as when the target person is single personnel, the behavior of the target person can be sleep, hair
It is slow-witted, work, steal etc.;Wherein, pilferage is non-permitted behavior.
Such as when the target person is at least two, the behavior of the target person is at least two target person
The combination behavior of member, such as can be talk, intimacy, fight etc..Wherein, fighting is non-permitted behavior.
It is understood which type of behavior is that non-permitted behavior can be configured as needed by manufacturer.
Wherein, existing Activity recognition technology can be used in Activity recognition technology, and the embodiment of the present invention also can be used and mention
The Activity recognition technology of confession.Activity recognition technology provided in an embodiment of the present invention be realized based on neural network model, such as
Can be convolutional neural networks (CNN:Convolutional Neural Network) model or Recognition with Recurrent Neural Network (RNN:
Recurrent Neural Networks) model, or it is also possible to deep neural network (DNN:Deep Neural
Networks) model.
104, the non-permitted behavior record is in the corresponding data base entries of identity of the target person.
Further, in one embodiment, it if the non-permitted behavior is hazardous act, such as steals, fights
Deng the robot can also alarm.
105, the behavior of the target person is not recorded.
From the above it can be seen that can carry out image using behavior monitoring method provided in an embodiment of the present invention to target person and adopt
Collection, to obtain the image of the target person, the image of the target person included at least the target person posture and
The face of the target person;Posture based on the target person identifies the image of the target person, thus really
The behavior of the fixed target person;Based on the target person face the image of the target person is identified, thus
Determine the identity of the target person;If the behavior of the target person is non-permitted behavior, the non-permitted behavior is remembered
Record is in the corresponding data base entries of the identity of the target person, to improve the efficiency of behavior monitoring;Further, exist
When the behavior of the user is hazardous act, it can alarm.
In one embodiment of the invention, the image recognition technology is deployed in neural network, neural network
It can be made of multiple neurons.In the neural network, the image recognition technology algorithm can be expressed as institute
The calculating formula stated:
M=f (pi+ λ)=f (Api+ λ), i=0 ..., n-1
Wherein, m indicates the facial characteristics identified, piIndicate the image of acquisition, the quantity of the image of acquisition is n, λ table
Show Dynamic gene, which has difference according to the image of different target persons, and what f () was indicated is that neuron is corresponding
Activation primitive, A is the corresponding module parameter of activation primitive.In one embodiment, activation primitive f () specifically can be
The form that sigmoid function, i.e. f () can be expressed as:
Wherein, the module parameter of activation primitive f () is trained in advance, and module parameter A has in one embodiment
Body can be by training function training to obtain as follows:
Wherein, M is the parameter of trained function, and N is the quantity of threedimensional model in training set, pnIt is target in training set
The picture of personnel, λnIt is the Dynamic gene in training set.
It is understood that the behavior to target person can also by a similar method, herein no longer when identifying
It repeats.
Fig. 2 describe it is provided by one embodiment of the present invention a kind of for carrying out the structure of the robot of behavior monitoring,
In, which can be used to implement the behavior monitoring method of previous embodiment offer.As shown in Fig. 2, the robot can wrap
It includes:
Image acquisition units 201, for carrying out Image Acquisition to target person, to obtain the image of the target person,
The image of the target person has included at least the posture of the target person and the face of the target person.
The target person refers to the personnel in workplace, such as office, meeting room, workshop etc..
In one embodiment, when carrying out Image Acquisition to target person, an image can be only acquired, to be based on
The image recognition technology of 2D carries out face recognition.It can complete to identify than faster in this way, to improve identification
Efficiency.
In another embodiment, in order to improve the accuracy of detection, the image recognition technology of 3D can be used, this
When, Image Acquisition is carried out to the target person from multiple angles, to obtain the more of the target person with different angle
A image.
Behavior determination unit 202 is known for image of the posture based on the target person to the target person
Not, with the behavior of the determination target person.
Wherein, in one embodiment, the image recognition technology of 2D is used, at this point it is possible to described in acquisition
The image of personnel carries out the extraction of facial characteristics, and the facial characteristics includes but is not limited to: the shape of face, between face
Distance etc..Then the face based on the target person identifies the image of the target person, so that it is determined that
The identity of the target person may include:
Face based on the target person extracts the facial characteristics of the target person;
It is searched from database according to the facial characteristics, so that it is determined that the identity of the target person.
In another embodiment, the image recognition technology of 3D is used, at this point, firstly the need of the personnel are based on
Multiple images to the personnel face carry out three-dimensional modeling, thus obtain the personnel face threedimensional model;Then
The facial characteristics of the personnel is determined based on the threedimensional model of the face of the personnel.At this point, the facial characteristics of the personnel removes
It may include the shape of face, it can also include the height of nose that the distance between face etc. are outer, the depth of eye socket, under
Bar convex-concave degree etc..It is described that Image Acquisition is carried out to target person, include: to obtain the image of the target person
Image Acquisition is carried out to the target person from multiple angles, to obtain the target person with different angle
Multiple images;
The face based on the target person identifies the image of the target person, so that it is determined that the mesh
The identity of mark personnel includes:
Multiple images based on the target person carry out three-dimensional modeling to the face of the target person, to obtain institute
State the threedimensional model of the face of target person;
The facial characteristics of the target person is determined based on the threedimensional model of the face of the target person;
It is searched from database according to the facial characteristics, so that it is determined that the identity of the target person.
Wherein, existing image recognition technology can be used in image recognition technology, and the embodiment of the present invention also can be used and mention
The image recognition technology of confession.Image recognition technology provided in an embodiment of the present invention be realized based on neural network model, such as
Can be convolutional neural networks (CNN:Convolutional Neural Network) model or Recognition with Recurrent Neural Network (RNN:
Recurrent Neural Networks) model, or it is also possible to deep neural network (DNN:Deep Neural
Networks) model.
Identity determination unit 203 knows the image of the target person for the face based on the target person
Not, with the identity of the determination target person.
Wherein, the posture based on the target person identifies the image of the target person, so that it is determined that
The behavior of the target person includes:
Posture based on the target person extracts the behavioural characteristic of the target person;
Based on the behavioural characteristic, the behavior of the target person is identified using pre-set algorithm.
Wherein, such as when the target person is single personnel, the behavior of the target person can be sleep, hair
It is slow-witted, work, steal etc.;Wherein, pilferage is non-permitted behavior.
Such as when the target person is at least two, the behavior of the target person is at least two target person
The combination behavior of member, such as can be talk, intimacy, fight etc..Wherein, fighting is non-permitted behavior.
It is understood which type of behavior is that non-permitted behavior can be configured as needed by manufacturer.
Wherein, existing Activity recognition technology can be used in Activity recognition technology, and the embodiment of the present invention also can be used and mention
The Activity recognition technology of confession.Activity recognition technology provided in an embodiment of the present invention be realized based on neural network model, such as
Can be convolutional neural networks (CNN:Convolutional Neural Network) model or Recognition with Recurrent Neural Network (RNN:
Recurrent Neural Networks) model, or it is also possible to deep neural network (DNN:Deep Neural
Networks) model.
Processing unit 204, for determining that the behavior of the target person is non-permitted row in the behavior determination unit 202
For when, then the non-permitted behavior record is to the identity pair of the target person that the identity determination unit 203 determines
In the data base entries answered.
Further, in one embodiment, it if the non-permitted behavior is hazardous act, such as steals, fights
Deng the processing unit 204 can also alarm.
Wherein, in one embodiment, the behavior determination unit 202 can be specifically used for:
Posture based on the target person extracts the behavioural characteristic of the target person;
Based on the behavioural characteristic, the behavior of the target person is identified using pre-set algorithm.
In one embodiment, the image recognition technology of 2D is used, at this point it is possible to the personnel's of acquisition
One image carries out the extraction of facial characteristics, and the facial characteristics includes but is not limited to: the shape of face, the distance between face
Etc..Then the identity determination unit 203 can be specifically used for:
Face based on the target person extracts the facial characteristics of the target person;
It is searched from database according to the facial characteristics, with the identity of the determination target person.
In another embodiment, the image recognition technology of 3D is used, at this point, firstly the need of the personnel are based on
Multiple images to the personnel face carry out three-dimensional modeling, thus obtain the personnel face threedimensional model;Then
The facial characteristics of the personnel is determined based on the threedimensional model of the face of the personnel.At this point, the facial characteristics of the personnel removes
It may include the shape of face, it can also include the height of nose that the distance between face etc. are outer, the depth of eye socket, under
Bar convex-concave degree etc..Then described image acquisition unit 201 can be specifically used for:
Image Acquisition is carried out to the target person from multiple angles, to obtain the target person with different angle
Multiple images;
The identity determination unit 203 can be specifically used for:
Multiple images based on the target person carry out three-dimensional modeling to the face of the target person, described in obtaining
The threedimensional model of the face of target person;
The facial characteristics of the target person is determined based on the threedimensional model of the face of the target person;
It is searched from database according to the facial characteristics, with the identity of the determination target person.
From the above it can be seen that using provided in an embodiment of the present invention for carrying out the robot of behavior monitoring, it can be to target person
Member carries out Image Acquisition, and to obtain the image of the target person, the image of the target person has included at least the target
The face of the posture of personnel and the target person;Posture based on the target person carries out the image of the target person
Identification, so that it is determined that the behavior of the target person;Face based on the target person to the image of the target person into
Row identification, so that it is determined that the identity of the target person;If the behavior of the target person is non-permitted behavior, will be described non-
Behavior record is allowed to be in the corresponding data base entries of the identity of the target person, to improve the efficiency of behavior monitoring;
Further, it when the behavior of the user is hazardous act, can alarm.
The embodiment of the invention also provides a kind of for carrying out the robot of behavior monitoring, can be used for executing aforementioned implementation
The behavior monitoring method that example provides.As shown in figure 3, the robot at least may include: memory 10, at least one processor
20, such as CPU (Central Processing Unit, central processing unit) and at least one image collecting device 30, it uses
In acquisition image.Wherein, between memory 10, processor 20 and image collecting device 30 can by one or more bus into
Row communication connection.The present invention is implemented it will be understood by those skilled in the art that the structure of robot shown in Fig. 3 is not constituted
The restriction of example, it is also possible to hub-and-spoke configuration either busbar network, can also include than illustrating more or fewer portions
Part perhaps combines certain components or different component layouts.
Wherein, memory 10 can be high speed RAM memory, be also possible to non-labile memory (non-
Volatile memory), a for example, at least magnetic disk storage.It is remote that memory 10 optionally can also be that at least one is located at
Storage device from aforementioned processor 20.Memory 10 can be used for storing executable program code and data, and the present invention is implemented
Example is not construed as limiting.
In robot for carrying out behavior monitoring shown in Fig. 3, processor 20 can be used for that memory 10 is called to deposit
The executable program code of storage executes following steps:
It controls image collecting device 30 and Image Acquisition is carried out to target person, to obtain the image of target person, target person
The image of member at least may include the posture of target person and the face of target person;
It is identified based on image of the posture of target person to target person, to determine the behavior of target person;
Face based on target person identifies the image of target person, to determine the identity of target person;
When the behavior for determining target person is non-permitted behavior, then the non-permitted behavior record is that memory 10 is deposited
In the corresponding data base entries of the identity of the target person of storage.
Optionally, processor 20 is identified based on image of the posture of target person to target person, to determine target
The mode of the behavior of personnel may include:
Posture based on target person extracts the behavioural characteristic of target person;
Based on behavior feature, the behavior of target person is identified using pre-set algorithm in memory 10.
Optionally, processor 20 identifies the image of target person based on the face of target person, to determine target
The mode of the identity of personnel may include:
Face based on target person extracts the facial characteristics of target person;
It is searched from the database that memory 10 stores according to the facial characteristics, to determine the identity of target person.
Optionally, processor 20 controls image collecting device 30 and carries out Image Acquisition to target person, to obtain target person
The mode of image of member may include:
It controls image collecting device 30 and Image Acquisition is carried out to target person from multiple angles, there is different angle to obtain
Target person multiple images;
Wherein, processor 20 identifies the image of target person based on the face of target person, to determine target person
The mode of identity of member may include:
Three-dimensional modeling is carried out based on face of the multiple images of target person to target person, to obtain the face of target person
The threedimensional model in portion;
The facial characteristics of target person is determined based on the threedimensional model of the face of target person;
It is searched from the database that memory 10 stores according to the facial characteristics, to determine the identity of target person.
Optionally, the quantity of target person can be two or more, the behavior of determining target person be it is above-mentioned at least
The combination behavior of two target persons.
Robot shown in implementing Fig. 3 can carry out Image Acquisition to target person, to obtain the figure of the target person
Picture, the image of the target person have included at least the posture of the target person and the face of the target person;Based on institute
The posture for stating target person identifies the image of the target person, so that it is determined that the behavior of the target person;It is based on
The face of the target person identifies the image of the target person, so that it is determined that the identity of the target person;If
The behavior of the target person is non-permitted behavior, then the non-permitted behavior record is to the identity pair of the target person
In the data base entries answered, to improve the efficiency of behavior monitoring;It further, is hazardous act in the behavior of the user
When, it can alarm.
The contents such as the information exchange between each unit module, implementation procedure that above-mentioned robot includes, due to the present invention
Embodiment of the method is based on same design, and for details, please refer to the description in the embodiment of the method for the present invention, and details are not described herein again.
The embodiment of the invention also provides a kind of computer readable storage medium, which has
Following steps may be implemented when being executed by processor in computer program, the computer program:
Image Acquisition is carried out to target person, to obtain the image of target person, the image of target person at least be can wrap
The posture of target person and the face of target person are included;
It is identified based on image of the posture of target person to target person, to determine the behavior of target person;
Face based on target person identifies the image of target person, to determine the identity of target person;
If the behavior of target person is non-permitted behavior, which is to the identity pair of target person
In the data base entries answered.
Optionally, processor is identified based on image of the posture of target person to target person, to determine target person
The mode of behavior of member may include:
Posture based on target person extracts the behavioural characteristic of target person;
Based on behavior feature, the behavior of target person is identified using algorithm pre-set in storage medium.
Optionally, processor identifies the image of target person based on the face of target person, to determine target person
The mode of identity of member may include:
Face based on target person extracts the facial characteristics of target person;
It is searched from database according to the facial characteristics, to determine the identity of target person.
Optionally, processor carries out Image Acquisition to target person, can wrap in a manner of the image for obtaining target person
It includes:
Image Acquisition is carried out to target person from multiple angles, to obtain multiple figures of the target person with different angle
Picture;
Wherein, processor identifies the image of target person based on the face of target person, to determine target person
The mode of identity may include:
Three-dimensional modeling is carried out based on face of the multiple images of target person to target person, to obtain the face of target person
The threedimensional model in portion;
The facial characteristics of target person is determined based on the threedimensional model of the face of target person;
It is searched from database according to the facial characteristics, to determine the identity of target person.
Optionally, the quantity of target person can be at least two, and the behavior of determining target person is above-mentioned at least two
The combination behavior of a target person.
Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, above-mentioned program can be stored in a computer-readable storage medium
In, the program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein, above-mentioned storage medium can be magnetic
Dish, CD, read-only memory (ROM:Read-Only Memory) or random access memory (RAM:Random
Access Memory) etc..
Used herein a specific example illustrates the principle and implementation of the invention, and above embodiments are said
It is bright to be merely used to help understand method and its thought of the invention;At the same time, for those skilled in the art, according to this hair
Bright thought, there will be changes in the specific implementation manner and application range, in conclusion the content of the present specification should not manage
Solution is limitation of the present invention.
Claims (10)
1. a kind of computer readable storage medium, which is characterized in that the storage medium is stored with computer program, the calculating
Machine program performs the steps of when being executed by processor
Image Acquisition is carried out to target person, to obtain the image of the target person, the image of the target person is at least wrapped
The posture of the target person and the face of the target person are included;
Posture based on the target person identifies the image of the target person, with the row of the determination target person
For;
Face based on the target person identifies the image of the target person, with the body of the determination target person
Part;
If the behavior of the target person is non-permitted behavior, the non-permitted behavior record is the target person
In the corresponding data base entries of identity.
2. computer readable storage medium as claimed in claim 1, which is characterized in that the processor is based on the target person
Posture identifies the image of the target person, includes: in a manner of the behavior of the determination target person
Posture based on the target person extracts the behavioural characteristic of the target person;
Based on the behavioural characteristic, the behavior of the target person is carried out using pre-set algorithm in the storage medium
Identification.
3. computer readable storage medium as claimed in claim 1 or 2, which is characterized in that the processor is based on the mesh
The face of mark personnel identifies the image of the target person, includes: in a manner of the identity of the determination target person
Face based on the target person extracts the facial characteristics of the target person;
It is searched from database according to the facial characteristics, with the identity of the determination target person.
4. computer readable storage medium as claimed in claim 1 or 2, which is characterized in that the processor is to target person
Image Acquisition is carried out, includes: in a manner of obtaining the image of the target person
Image Acquisition is carried out to the target person from multiple angles, to obtain the more of the target person with different angle
A image;
Wherein, the processor identifies the image of the target person based on the face of the target person, with determination
The mode of the identity of the target person includes:
Multiple images based on the target person carry out three-dimensional modeling to the face of the target person, to obtain the target
The threedimensional model of the face of personnel;
The facial characteristics of the target person is determined based on the threedimensional model of the face of the target person;
It is searched from database according to the facial characteristics, with the identity of the determination target person.
5. computer readable storage medium according to claim 1 or 2, which is characterized in that the quantity of the target person
It is at least two, the behavior of the target person of the determination is the combination behavior of at least two target person.
6. a kind of for carrying out the robot of behavior monitoring characterized by comprising memory, processor and image collector
It sets, wherein the memory is used to acquire image for storing executable program code and data, described image acquisition device,
The processor is used to call the executable program code of the memory storage, executes following steps:
It controls described image acquisition device and Image Acquisition is carried out to target person, it is described to obtain the image of the target person
The image of target person has included at least the posture of the target person and the face of the target person;
Posture based on the target person identifies the image of the target person, with the row of the determination target person
For;
Face based on the target person identifies the image of the target person, with the body of the determination target person
Part;
When the behavior for determining the target person is non-permitted behavior, then the non-permitted behavior record is the storage
In the corresponding data base entries of identity of the target person of device storage.
7. robot as claimed in claim 6, which is characterized in that posture of the processor based on the target person is to the mesh
The image of mark personnel identifies, includes: in a manner of the behavior of the determination target person
Posture based on the target person extracts the behavioural characteristic of the target person;
Based on the behavioural characteristic, the behavior of the target person is known using pre-set algorithm in the memory
Not.
8. robot as claimed in claims 6 or 7, which is characterized in that face of the processor based on the target person
The image of the target person is identified, includes: in a manner of the identity of the determination target person
Face based on the target person extracts the facial characteristics of the target person;
It is searched from the database that the memory stores according to the facial characteristics, with the body of the determination target person
Part.
9. robot as claimed in claims 6 or 7, which is characterized in that the processor controls described image acquisition device pair
Target person carries out Image Acquisition, includes: in a manner of obtaining the image of the target person
It controls described image acquisition device and Image Acquisition is carried out to the target person from multiple angles, there are different angles to obtain
The multiple images of the target person of degree;
Wherein, the processor identifies the image of the target person based on the face of the target person, with determination
The mode of the identity of the target person includes:
Multiple images based on the target person carry out three-dimensional modeling to the face of the target person, to obtain the target
The threedimensional model of the face of personnel;
The facial characteristics of the target person is determined based on the threedimensional model of the face of the target person;
It is searched from the database that the memory stores according to the facial characteristics, with the body of the determination target person
Part.
10. robot as claimed in claims 6 or 7, which is characterized in that the quantity of the target person be it is two or more,
The behavior of the target person of the determination is the combination behavior of at least two target person.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810927706.0A CN109086731A (en) | 2018-08-15 | 2018-08-15 | It is a kind of for carrying out the robot and storage medium of behavior monitoring |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810927706.0A CN109086731A (en) | 2018-08-15 | 2018-08-15 | It is a kind of for carrying out the robot and storage medium of behavior monitoring |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109086731A true CN109086731A (en) | 2018-12-25 |
Family
ID=64793487
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810927706.0A Pending CN109086731A (en) | 2018-08-15 | 2018-08-15 | It is a kind of for carrying out the robot and storage medium of behavior monitoring |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109086731A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109495727A (en) * | 2019-01-04 | 2019-03-19 | 京东方科技集团股份有限公司 | Intelligent control method and device, system, readable storage medium storing program for executing |
CN110855797A (en) * | 2019-11-28 | 2020-02-28 | 河北农业大学 | Sheep behavior monitoring system and method based on action recognition |
CN111611904A (en) * | 2020-05-15 | 2020-09-01 | 新石器慧通(北京)科技有限公司 | Dynamic target identification method based on unmanned vehicle driving process |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104581047A (en) * | 2014-12-15 | 2015-04-29 | 苏州福丰科技有限公司 | Three-dimensional face recognition method for supervisory video recording |
CN104615983A (en) * | 2015-01-28 | 2015-05-13 | 中国科学院自动化研究所 | Behavior identification method based on recurrent neural network and human skeleton movement sequences |
CN105718857A (en) * | 2016-01-13 | 2016-06-29 | 兴唐通信科技有限公司 | Human body abnormal behavior detection method and system |
WO2016109841A1 (en) * | 2014-12-31 | 2016-07-07 | Morphotrust Usa, Llc | Detecting facial liveliness |
CN205540806U (en) * | 2016-04-07 | 2016-08-31 | 小煷伴(深圳)智能科技有限公司 | Three -dimensional facial image recognition device |
CN106780569A (en) * | 2016-11-18 | 2017-05-31 | 深圳市唯特视科技有限公司 | A kind of human body attitude estimates behavior analysis method |
CN107103310A (en) * | 2017-06-01 | 2017-08-29 | 鄂尔多斯市普渡科技有限公司 | The passenger behavior monitor system and method for a kind of unmanned taxi |
CN107609517A (en) * | 2017-09-15 | 2018-01-19 | 华中科技大学 | A kind of classroom behavior detecting system based on computer vision |
CN107644190A (en) * | 2016-07-20 | 2018-01-30 | 北京旷视科技有限公司 | Pedestrian's monitoring method and device |
-
2018
- 2018-08-15 CN CN201810927706.0A patent/CN109086731A/en active Pending
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104581047A (en) * | 2014-12-15 | 2015-04-29 | 苏州福丰科技有限公司 | Three-dimensional face recognition method for supervisory video recording |
WO2016109841A1 (en) * | 2014-12-31 | 2016-07-07 | Morphotrust Usa, Llc | Detecting facial liveliness |
CN104615983A (en) * | 2015-01-28 | 2015-05-13 | 中国科学院自动化研究所 | Behavior identification method based on recurrent neural network and human skeleton movement sequences |
CN105718857A (en) * | 2016-01-13 | 2016-06-29 | 兴唐通信科技有限公司 | Human body abnormal behavior detection method and system |
CN205540806U (en) * | 2016-04-07 | 2016-08-31 | 小煷伴(深圳)智能科技有限公司 | Three -dimensional facial image recognition device |
CN107644190A (en) * | 2016-07-20 | 2018-01-30 | 北京旷视科技有限公司 | Pedestrian's monitoring method and device |
CN106780569A (en) * | 2016-11-18 | 2017-05-31 | 深圳市唯特视科技有限公司 | A kind of human body attitude estimates behavior analysis method |
CN107103310A (en) * | 2017-06-01 | 2017-08-29 | 鄂尔多斯市普渡科技有限公司 | The passenger behavior monitor system and method for a kind of unmanned taxi |
CN107609517A (en) * | 2017-09-15 | 2018-01-19 | 华中科技大学 | A kind of classroom behavior detecting system based on computer vision |
Non-Patent Citations (1)
Title |
---|
杨秀金: "仓库值班员行为视频智能监控系统设计与开发", 《电脑编程技巧与维护》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109495727A (en) * | 2019-01-04 | 2019-03-19 | 京东方科技集团股份有限公司 | Intelligent control method and device, system, readable storage medium storing program for executing |
CN110855797A (en) * | 2019-11-28 | 2020-02-28 | 河北农业大学 | Sheep behavior monitoring system and method based on action recognition |
CN111611904A (en) * | 2020-05-15 | 2020-09-01 | 新石器慧通(北京)科技有限公司 | Dynamic target identification method based on unmanned vehicle driving process |
CN111611904B (en) * | 2020-05-15 | 2023-12-01 | 新石器慧通(北京)科技有限公司 | Dynamic target identification method based on unmanned vehicle driving process |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Wayne et al. | Unsupervised predictive memory in a goal-directed agent | |
CN108759844B (en) | Robot repositioning and environment map constructing method, robot and storage medium | |
CN111985385B (en) | Behavior detection method, device and equipment | |
CN109086731A (en) | It is a kind of for carrying out the robot and storage medium of behavior monitoring | |
CN106378780A (en) | Robot system and method and server for controlling robot | |
CN107967513B (en) | Multirobot intensified learning collaboratively searching method and system | |
CN109085833A (en) | A kind of patrol robot and storage medium | |
CN109101928A (en) | WatchChild method and robot | |
Brom et al. | Episodic memory for human-like agents and human-like agents for episodic memory | |
CN109003329A (en) | A kind of target goods heap monitoring device and storage medium | |
CN109118580A (en) | Target goods heap monitoring method and relevant apparatus | |
CN109191633A (en) | Employee's recognition methods and robot | |
Jayaratne et al. | Bio-inspired multisensory fusion for autonomous robots | |
CN109088452B (en) | Robot charging method and robot | |
CN108858201A (en) | It is a kind of for nursing the robot and storage medium of children | |
CN109145804A (en) | Behavior monitoring method and robot | |
CN113642720A (en) | Bionic scene cognition model construction method and system | |
CN108981712A (en) | Robot goes on patrol method and robot | |
Linton et al. | New approaches to 3D vision | |
Sanghvi et al. | Mgpi: A computational model of multiagent group perception and interaction | |
Tan | Implementation of a framework for imitation learning on a humanoid robot using a cognitive architecture | |
US20220305647A1 (en) | Future prediction, using stochastic adversarial based sampling, for robotic control and/or other purpose(s) | |
CN108877115A (en) | Evacuate guidance method and robot | |
CN109141454A (en) | A kind of charging robot and storage medium | |
CN109048899A (en) | A kind of patrol robot and storage medium |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20181225 |