CN109327328A - Monitoring and managing method, device, system, cloud server and storage medium - Google Patents
Monitoring and managing method, device, system, cloud server and storage medium Download PDFInfo
- Publication number
- CN109327328A CN109327328A CN201810982884.3A CN201810982884A CN109327328A CN 109327328 A CN109327328 A CN 109327328A CN 201810982884 A CN201810982884 A CN 201810982884A CN 109327328 A CN109327328 A CN 109327328A
- Authority
- CN
- China
- Prior art keywords
- monitoring information
- monitoring
- information
- suspicious
- anomalous event
- 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L63/00—Network architectures or network communication protocols for network security
- H04L63/14—Network architectures or network communication protocols for network security for detecting or protecting against malicious traffic
- H04L63/1408—Network architectures or network communication protocols for network security for detecting or protecting against malicious traffic by monitoring network traffic
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L63/00—Network architectures or network communication protocols for network security
- H04L63/14—Network architectures or network communication protocols for network security for detecting or protecting against malicious traffic
- H04L63/1408—Network architectures or network communication protocols for network security for detecting or protecting against malicious traffic by monitoring network traffic
- H04L63/1425—Traffic logging, e.g. anomaly detection
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
Abstract
This disclosure relates to a kind of monitoring and managing method, device, system, cloud server and storage medium, for realizing to the supervision of application services and the automatic evidence-collectings of anomalous event such as nets about vehicle, family's monitoring, online live streaming platform.The method is applied to cloud server, the described method includes: receiving the suspicious monitoring information that monitoring client is sent, the suspicious monitoring information is that the monitoring client identifies that the obtained monitoring information with anomalous event, the list feature identification model are obtained using the single features in monitoring information as training sample training based on single feature identification model from collected monitoring information；The suspicious monitoring information is inputted into multiple features fusion model, obtains the recognition result of anomalous event, the multiple features fusion model is to obtain using the anomalous event label of multiple features and the monitoring information in monitoring information as training sample to training.
Description
Technical field
This disclosure relates to field of artificial intelligence, and in particular, to a kind of monitoring and managing method, device, system, cloud service
Device and storage medium.
Background technique
With the continuous development of internet technique, the application services such as net about vehicle, online live streaming platform are come into being, at this
It is possible that violence, quarrelling, relating to all kinds of unlawful practices such as yellow, vulgar during the operation of a little application services.However, mesh
It is preceding the operation process of these application services not to be supervised, cause all kinds of delinquent events to emerge one after another.
Summary of the invention
Purpose of this disclosure is to provide a kind of monitoring and managing method, device, system, cloud server and storage mediums, for real
Now to the supervision of application services and the automatic evidence-collectings of anomalous event such as nets about vehicle, family's monitoring, online live streaming platform.
To achieve the goals above, disclosure first aspect provides a kind of monitoring and managing method, is applied to cloud server, described
Method includes:
The suspicious monitoring information that monitoring client is sent is received, the suspicious monitoring information is that the monitoring client is based on single feature knowledge
Other model identifies the obtained monitoring information with anomalous event, the list feature identification model from collected monitoring information
It is to be obtained using the single features in monitoring information as training sample training；
The suspicious monitoring information is inputted into multiple features fusion model, obtains the recognition result of anomalous event, it is described mostly special
Levying Fusion Model is using the anomalous event label of multiple features and the monitoring information in monitoring information as training sample
Training is obtained.
Disclosure second aspect provides a kind of maintenance device, is applied to cloud server, and described device includes:
First receiving module is configured as receiving the suspicious monitoring information that monitoring client is sent, and the suspicious monitoring information is
The monitoring client identifies the obtained monitoring with anomalous event based on single feature identification model from collected monitoring information
Information, the list feature identification model are obtained using the single features in monitoring information as training sample training；
Identification module is configured as the suspicious monitoring information input multiple features fusion model obtaining anomalous event
Recognition result, the multiple features fusion model are the anomalous events using multiple features and the monitoring information in monitoring information
Label obtains training as training sample.
The disclosure third aspect provides a kind of computer readable storage medium, is stored thereon with computer program, the program
The step of disclosure first aspect the method is realized when being executed by processor.
Disclosure fourth aspect provides a kind of cloud server, comprising:
Memory is stored thereon with computer program；
Processor, for executing the computer program in the memory, to realize described in disclosure first aspect
The step of method.
The 5th aspect of the disclosure provides a kind of supervisory systems, takes including cloud described in monitoring client and disclosure fourth aspect
Business device.
By adopting the above technical scheme, it at least can achieve following technical effect:
By receiving the suspicious monitoring information of monitoring client transmission and suspicious monitoring information being inputted multiple features fusion model, obtain
To the recognition result of anomalous event, may be implemented supervision to application services such as nets about vehicle, family's monitoring, online live streaming platform and
The automatic evidence-collecting of anomalous event reduces the probability that anomalous event occurs.Secondly, being carried out by monitoring information of the monitoring client to acquisition
The suspicious monitoring information that identification obtains simultaneously is sent to cloud server by preliminary identification, by cloud server to suspicious monitoring information
It is further identified, on the one hand can reduce the operand of monitoring client, and then reduce the bandwidth of monitoring client, it on the other hand can be with
The accuracy and reliability for improving anomalous event recognition result reduces the false alarm rate of anomalous event, and then reduces human cost.
Other feature and advantage of the disclosure will the following detailed description will be given in the detailed implementation section.
Detailed description of the invention
In order to illustrate more clearly of the embodiment of the present disclosure or technical solution in the prior art, embodiment will be described below
Needed in attached drawing be briefly described, it should be apparent that, the accompanying drawings in the following description is some realities of the disclosure
Example is applied, it for those of ordinary skill in the art, without creative efforts, can also be according to these attached drawings
Obtain other attached drawings.
Fig. 1 is the block diagram according to a kind of supervisory systems shown in one exemplary embodiment of the disclosure；
Fig. 2 is the flow chart according to a kind of monitoring and managing method shown in one exemplary embodiment of the disclosure；
Fig. 3 is a kind of flow chart of monitoring and managing method shown according to disclosure another exemplary embodiment；
Fig. 4 is the block diagram according to a kind of maintenance device shown in one exemplary embodiment of the disclosure；
Fig. 5 is a kind of block diagram of maintenance device shown according to disclosure another exemplary embodiment；
Fig. 6 is the structural schematic diagram according to a kind of cloud server shown in one exemplary embodiment of the disclosure.
Specific embodiment
To keep the purposes, technical schemes and advantages of the embodiment of the present disclosure clearer, below in conjunction with the embodiment of the present disclosure
In attached drawing, the technical solution in the embodiment of the present disclosure is clearly and completely described, it is clear that described embodiment is
Disclosure a part of the embodiment, instead of all the embodiments.Based on the embodiment in the disclosure, those of ordinary skill in the art
Every other embodiment obtained without creative efforts belongs to the range of disclosure protection.
It should be noted that the specification and claims of the disclosure and term " first " in above-mentioned attached drawing, "
Two " etc. be to be used to distinguish similar objects, and describes specific sequence or precedence without being interpreted as.
The embodiment of the present disclosure provides a kind of supervisory systems, as shown in Figure 1, the supervisory systems 100 may include: to be located at now
The monitoring client 110 and cloud server 120 of field.Monitoring client 110 can be to have merged information collection and the intelligence of information processing is set
It is standby, it may include information collecting device 111 and information processing unit 112.Information collecting device 111 will after acquiring monitoring information
Monitoring information is sent to information processing unit 112 and is identified, tentatively judge whether there is anomalous event (such as violence, rape,
Quarrel, relate to yellow, vulgar etc.), and obtained suspicious monitoring information (i.e. with the monitoring information of anomalous event) will be identified by having
The mode of line (such as cable, cable) or wireless (such as WiFi, bluetooth) are sent to cloud server 120.Cloud server 120
It is further identified according to the suspicious monitoring information received, obtains the recognition result of anomalous event (for example including monitoring client
110 with the presence or absence of anomalous event and there are which kind of anomalous events).
Wherein, information collecting device 111 can include but is not limited to microphone, camera, GPS positioning components, inertia survey
Sensors such as unit (Inertial Measurement Unite, IMU) etc. are measured, correspondingly, collected monitoring information can be with
Including but not limited to: voice messaging, image information, the movement state information of text information, the location information of vehicle and vehicle
Deng.
Information processing unit 112 is built-in with single feature identification model, which is using in monitoring information
Single features obtained as training sample training.Correspondingly, information processing unit 112 can identify mould based on the list feature
Type is identified from collected monitoring information obtains suspicious monitoring information.Specifically, which may include needle
To the feature extraction network and classifier of each feature, information processing unit 112 can use each feature extraction network from monitoring
Corresponding feature is extracted in information and is classified using corresponding classifier to this feature, tentatively judges that there is abnormal thing
The suspicious monitoring information of part.For example, if identifying it according to single feature identification model, there are abnormal sound (examples for voice messaging
Such as shriek), only have vehicle-mounted sound (i.e. microphone in abnormal keyword (such as " releasing me ", " help ") and voice messaging
At least one of it is blocked) these three situations, then it is believed that the voice messaging is suspicious monitoring information；For image information,
If recognizing dangerous article (such as knife, rifle etc.), image information in image information according to single feature identification model to show to supervise
It controls end and strenuous exercise and image information occurs as in abnormal image (such as image of non-monitored object) these three situations
At least one, the then it is believed that image information is suspicious monitoring information；For text information, if being known according to single feature identification model
Being clipped to includes illegal text (such as violence, relate to the texts such as yellow, vulgar) in text information, then it is believed that text information is suspicious
Text information.
Cloud server 120 is built-in with multiple features fusion model, and the multiple features fusion model is using in monitoring information
The anomalous event label of multiple features and the monitoring information obtains training as training sample.Cloud server 120 exists
After the suspicious monitoring information for receiving the transmission of information processing unit 112, suspicious monitoring information can be inputted to preset multiple features
In Fusion Model, the recognition result of anomalous event is obtained.Specifically, which may include for each feature
Feature extraction network, multimodality fusion network and classifier for all features extracted, cloud server 120 can benefit
Multiple features are extracted from suspicious monitoring information with each feature extraction network, using multimodality fusion network by multiple features into
Row fusion obtains feature vector, and feature vector is sent into classifier and carries out Classification and Identification, obtains the identification knot of anomalous event
Fruit.
By the embodiment of the present disclosure provide supervisory systems, may be implemented to net about vehicle, family monitoring, platform is broadcast live online
The supervision of equal application services and the automatic evidence-collecting of anomalous event, reduce the probability that anomalous event occurs.Secondly, passing through monitoring client pair
The monitoring information of acquisition is tentatively identified and the suspicious monitoring information that identification obtains is sent to cloud server, is taken by cloud
Business device further identifies suspicious monitoring information using multiple features fusion model, on the one hand can reduce the operation of monitoring client
Amount, and then the bandwidth of monitoring client is reduced, the accuracy and reliability of anomalous event recognition result on the other hand can be improved, in turn
The false alarm rate of anomalous event is reduced, human cost is reduced.
In another embodiment, if recognition result shows monitoring client 110, there are anomalous event, cloud servers 120
Suspicious monitoring information can be sent to destination 200 (for example, remote supervisory and control(ling) equipment, the portable movement of target user are eventually
End etc.), which is shown with indicative purpose end 200, so that target user is further according to the suspicious monitoring information
Monitoring client 110 is identified with the presence or absence of anomalous event and inputs manual identified result to destination 200 and is confirming monitoring client 110
There are take appropriate measures when anomalous event (such as alarms).Wherein, target user can include but is not limited to monitoring personnel,
Network police, contact staff of application service etc..In this way, by monitoring client, cloud server and artificial three-level recognition mechanism, it can
Further to promote the accuracy and reliability of anomalous event recognition result, anomalous event false alarm rate is reduced, to reduce artificial
Cost.Further, since destination only shows the suspicious monitoring information after monitoring client screens, all adopt is shown compared to destination
The monitoring information of collection can guarantee the privacy of monitored object to a certain extent, realize monitoring client bandwidth, be monitored object
Optimization compromise between privacy and anomalous event discrimination.
Further, cloud server 120 can also receive the artificial knowledge to suspicious monitoring information of the transmission of destination 200
Not as a result, and training sample pair is generated according to suspicious monitoring information and manual identified result, it is more to updating using the training sample
Fusion Features model promotes multiple features fusion model according to suspicious prison to optimize and be promoted the performance of the multiple features fusion model
The accuracy and reliability of the recognition result of control information output reaches labor-saving to reduce the false alarm rate of anomalous event
Purpose.Meanwhile as training sample increases quantity, by constantly updating and optimizing, multiple features fusion model can be made more
It adds kind.
Further, the manual identified result received can be also sent to the information of monitoring client 110 by cloud server 120
Processing unit 112 generates training according to collected monitoring information and corresponding manual identified result by information processing unit 112
Sample to and using the training sample that generates to single feature identification model is updated, to optimize and be promoted the list feature identification model
Performance, and then promote single feature identification model and the accuracy of suspicious monitoring information and reliable is exported according to collected monitoring information
Property, to reduce the false alarm rate of anomalous event.Meanwhile increasing with training samples number can by constantly updating and optimizing
So that single feature identification model is more perfect.
It is worth noting that the application scenarios of the embodiment of the present disclosure can include but is not limited to: the embodiment of the present disclosure is answered
It can include but is not limited to scene: net about vehicle supervision, family's monitoring, online live streaming platform supervision etc..
In addition, being directed to different application scenarios, the monitoring information that monitoring client 110 acquires is different.For example, corresponding net about vehicle is supervised
Pipe, monitoring client 110 can be set on vehicle, it is contemplated that rape, murder etc. anomalous events mostly occur in isolated area and
Vehicle is in halted state mostly, thus the monitoring information that monitoring client 110 acquires may include: voice messaging, image information, vehicle
Location information and vehicle movement state information；For platform supervision is broadcast live online, monitoring client 110 be can be set straight
It broadcasts in client, it is contemplated that there is text during live streaming, thus the monitoring information that monitoring client 110 acquires may include: voice letter
Breath, image information and text information.
It will be understood by those skilled in the art that the embodiment of the present disclosure is only with the above-mentioned application scenarios of exemplary illustration, but not
It is to limit the application range of the embodiment of the present disclosure.
The embodiment of the present disclosure additionally provides a kind of monitoring and managing method, as shown in Fig. 2, this method may comprise steps of:
In the step s 21, the suspicious monitoring information that monitoring client is sent is received, suspicious monitoring information is that monitoring client is based on Dan Te
Sign identification model identifies the obtained monitoring information with anomalous event from collected monitoring information.
Wherein, single feature identification model is obtained using the single features in monitoring information as training sample training.
According to the difference of application scenarios, monitoring information may include image information, voice messaging, text information, position letter
One or more of breath and movement state information.Illustratively, by taking net about vehicle supervision scene as an example, it is contemplated that rape, murder
Etc. anomalous events mostly occur in isolated area and vehicle is in halted state mostly, thus available monitoring information is removed
Include image information and voice messaging, can also include the location information of vehicle and the movement state information of vehicle.With
Line is broadcast live for platform supervision scene, it is contemplated that have a text during live streaming, it is available to monitoring information include image letter
Breath and voice messaging, can also include text information.
It is worth noting that the tool that monitoring client is identified from collected monitoring information using single feature identification model
There is the monitoring client in supervisory systems described in the process and any of the above-described a embodiment of the disclosure of the monitoring information of anomalous event
Specific implementation it is similar, specifically refer to the above-mentioned description to monitoring client, in order to reduce redundancy, details are not described herein again.
In step S22, suspicious monitoring information is inputted into multiple features fusion model, obtains the recognition result of anomalous event.
Wherein, multiple features fusion model is the anomalous event using multiple features and the monitoring information in monitoring information
Label obtains training as training sample.
Specifically, multiple features fusion model may include for the feature extraction network of each feature, for all extractions
The multimodality fusion network and classifier that the multiple features of feature out are merged.After receiving suspicious information, it can use each
A feature extraction network extracts multiple features from suspicious monitoring information, is melted multiple features using multimodality fusion network
It closes, obtains feature vector, and feature vector is sent into classifier and carries out Classification and Identification, obtain the recognition result of anomalous event.
In embodiment of the disclosure, the recognition result of anomalous event may include monitoring client with the presence or absence of anomalous event
(such as violence, rape, quarrel, relating to yellow, vulgar etc.) and there are which kind of anomalous events.
According to disclosure monitoring and managing method provided by the above embodiment, may be implemented to net about vehicle, family's monitoring, online live streaming
The supervision of the application services such as platform and the automatic evidence-collecting of anomalous event reduce the probability that anomalous event occurs.Secondly, passing through monitoring
End tentatively identifies the monitoring information of acquisition and the suspicious monitoring information that identification obtains is sent to cloud server, by cloud
End server by utilizing multiple features fusion model further identifies suspicious monitoring information, on the one hand can reduce monitoring client
Operand, and then the bandwidth of monitoring client is reduced, the accuracy and reliability of anomalous event recognition result on the other hand can be improved,
And then the false alarm rate of anomalous event is reduced, reduce human cost.
In another embodiment, as shown in figure 3, above-mentioned monitoring and managing method can also include:
In step S23, if recognition result shows monitoring client, there are anomalous events, and suspicious monitoring information is sent to mesh
End, suspicious monitoring information is shown with indicative purpose end.
Wherein, destination can include but is not limited to remote supervisory and control(ling) equipment, mobile terminal that target user carries etc., target
User can for example including but be not limited to monitoring personnel, network police, the contact staff of application service etc..
If recognition result show monitoring client there are anomalous event, cloud server can also by suspicious monitoring information (such as
Suspicious voice messaging, image information etc.) it is sent to destination, prison is further identified according to suspicious monitoring information by target user
Controlling end whether there is anomalous event, so that in confirmation monitoring client, there are take appropriate measures when anomalous event (such as alarms).
As a result, by monitoring client, cloud server and artificial three-level recognition mechanism, anomalous event knowledge can be further promoted
The accuracy and reliability of other result reduces anomalous event false alarm rate, to be further reduced cost of labor.Further, since mesh
End only show the suspicious monitoring information after monitoring client screens, can compared to the monitoring information that destination shows all acquisitions
To guarantee the privacy of monitored object to a certain extent, monitoring client bandwidth, the privacy of monitored object and abnormal thing are realized
Optimization compromise between part discrimination.
In another embodiment, as shown in figure 3, above-mentioned monitoring and managing method can also include:
In step s 24, the manual identified result to suspicious monitoring information that destination is sent is received.
In step s 25, training sample pair is generated according to suspicious monitoring information and manual identified result.
In step S26, using the training sample to update multiple features fusion model.
Target user can also input to the manual identified of suspicious monitoring information in destination as a result, will be artificial by destination
Recognition result is sent to cloud server.Cloud server is after receiving manual identified result, to suspicious monitoring information and right
The manual identified result answered is labeled, and obtains training sample pair, and using the training sample to updating multiple features fusion model,
To optimize and be promoted the performance of the multiple features fusion model, the knowledge that multiple features fusion model is exported according to suspicious monitoring information is promoted
The accuracy and reliability of other result reaches labor-saving purpose to reduce the false alarm rate of anomalous event.Meanwhile with
Training sample increases quantity, by constantly updating and optimizing, multiple features fusion model can be made more perfect.
In another embodiment, as shown in figure 3, above-mentioned monitoring and managing method can also include:
In step s 27, manual identified result is sent to monitoring client, manual identified result updates Dan Te for monitoring client
Levy identification model.
The manual identified result received can also be sent to monitoring client by cloud server, by monitoring client according to collecting
Monitoring information and corresponding manual identified result generate training sample to and using the training sample that generates to updating single feature
Identification model to optimize and be promoted the performance of the list feature identification model, and then promotes single feature identification model according to collecting
Monitoring information export the accuracy and reliability of suspicious monitoring information, to reduce the false alarm rate of anomalous event.Meanwhile with
Training samples number increases, and by constantly updating and optimizing, single feature identification model can be made more perfect.
It is worth noting that for simple description, therefore, it is stated as a series of dynamic for above method embodiment
It combines, but those skilled in the art should understand that, the present invention is not limited by the sequence of acts described.Secondly, this
Field technical staff also should be aware of, and the embodiments described in the specification are all preferred embodiments, and related movement is simultaneously
It is not necessarily necessary to the present invention.
The embodiment of the present disclosure also provides a kind of maintenance device, which can be applied to cloud server, as shown in figure 4, should
Maintenance device 400 may include: the first receiving module 401 and identification module 402.
First receiving module 401 is configured as receiving the suspicious monitoring information that monitoring client is sent, the suspicious monitoring letter
Breath is that the monitoring client identifies that is obtained has anomalous event based on single feature identification model from collected monitoring information
Monitoring information, the list feature identification model are obtained using the single features in monitoring information as training sample training.
The identification module 402 is configured as the suspicious monitoring information input multiple features fusion model obtaining abnormal thing
The recognition result of part, the multiple features fusion model are the exceptions using multiple features and the monitoring information in monitoring information
Event tag obtains training as training sample.
In another embodiment, as shown in figure 5, described device 400 further include:
First sending module 403 is configured as showing that the monitoring client there are when anomalous event, is incited somebody to action in the recognition result
The suspicious monitoring information is sent to destination, to indicate that the destination shows the suspicious monitoring information.
In another embodiment, as shown in figure 5, described device 400 further include:
Second receiving module 404, is configured as receiving that the destination sends to the artificial of the suspicious monitoring information
Recognition result；
Generation module 405 is configured as generating training sample according to the suspicious monitoring information and the manual identified result
This is right；
Update module 406 is configured as using the training sample to the update multiple features fusion model.
In another embodiment, as shown in figure 5, described device 400 further include:
Second sending module 407 is configured as the manual identified result being sent to the monitoring client, the artificial knowledge
Other result updates single feature identification model for the monitoring client.
In another embodiment, the monitoring information includes image information, voice messaging, text information, location information
And one or more of movement state information.
Those skilled in the art can be understood that, for convenience and simplicity of description, only with above-mentioned each function mould
The division progress of block can according to need and for example, in practical application by above-mentioned function distribution by different functional modules
It completes, i.e., the internal structure of device is divided into different functional modules, to complete all or part of the functions described above.
The specific work process of foregoing description functional module, can refer to corresponding processes in the foregoing method embodiment, no longer superfluous herein
It states.
By disclosure maintenance device provided by the above embodiment, may be implemented to net about vehicle, family's monitoring, online live streaming
The supervision of the application services such as platform and the automatic evidence-collecting of anomalous event reduce the probability that anomalous event occurs.Secondly, passing through monitoring
End tentatively identifies the monitoring information of acquisition and the suspicious monitoring information that identification obtains is sent to cloud server, by cloud
End server by utilizing multiple features fusion model further identifies suspicious monitoring information, on the one hand can reduce monitoring client
Operand, and then the bandwidth of monitoring client is reduced, the accuracy and reliability of anomalous event recognition result on the other hand can be improved,
And then the false alarm rate of anomalous event is reduced, reduce human cost.
Secondly, show that suspicious monitoring information there are when anomalous event, is sent to destination by monitoring client in recognition result, with
Indicative purpose end shows suspicious monitoring information, further identifies that monitoring client whether there is according to suspicious monitoring information by target user
Anomalous event, can be into one by monitoring client, cloud server and artificial three-level recognition mechanism to take appropriate measures
Step promotes the accuracy and reliability of anomalous event recognition result, reduces anomalous event false alarm rate, to be further reduced artificial
Cost.Further, since destination only shows the suspicious monitoring information after monitoring client screens, all adopt is shown compared to destination
The monitoring information of collection can guarantee the privacy of monitored object to a certain extent, realize monitoring client bandwidth, be monitored object
Optimization compromise between privacy and anomalous event discrimination.
Further, multiple features fusion model is updated using the manual identified result received and suspicious monitoring information, it can
To optimize and be promoted the performance of the multiple features fusion model, the knowledge that multiple features fusion model is exported according to suspicious monitoring information is promoted
The accuracy and reliability of other result reaches labor-saving purpose to reduce the false alarm rate of anomalous event.Meanwhile with
Training sample increases quantity, by constantly updating and optimizing, multiple features fusion model can be made more perfect.
Further, monitoring client is sent to by the manual identified result that will be received, manual identified is utilized by monitoring client
As a result single feature identification model is updated, the performance of the list feature identification model can be optimized and be promoted, and then promotes single feature and knows
Other model exports the accuracy and reliability of suspicious monitoring information according to collected monitoring information, to reduce anomalous event
False alarm rate.Meanwhile increasing with training samples number can make single feature identification model more by constantly updating and optimizing
It adds kind.
Fig. 6 is a kind of block diagram of cloud server 600 shown according to an exemplary embodiment.Referring to Fig. 6, cloud service
Device 600 includes processor 622, and quantity can be one or more and memory 632, can be by processor 622 for storing
The computer program of execution.The computer program stored in memory 632 may include that one or more each is right
The module of Ying Yuyi group instruction.In addition, processor 622, which can be configured as, executes the computer program, to execute above-mentioned prison
Pipe method.
In addition, cloud server 600 can also include power supply module 626 and communication component 650, which can
To be configured as executing the power management of cloud server 600, which can be configured as realization cloud server
600 communication, for example, wired or wireless communication.In addition, the cloud server 600 can also include that input/output (I/O) connects
Mouth 658.Cloud server 600 can be operated based on the operating system for being stored in memory 632, such as Windows
ServerTM, Mac OS XTM, UnixTM, LinuxTM etc..
In a further exemplary embodiment, a kind of computer readable storage medium including program instruction is additionally provided, it should
The step of above-mentioned monitoring and managing method is realized when program instruction is executed by processor.For example, the computer readable storage medium can be with
For the above-mentioned memory 632 including program instruction, above procedure instruction can be executed by the processor 622 of cloud server 600 with
Complete above-mentioned monitoring and managing method.
The preferred embodiment of the disclosure is described in detail in conjunction with attached drawing above, still, the disclosure is not limited to above-mentioned reality
The detail in mode is applied, in the range of the technology design of the disclosure, a variety of letters can be carried out to the technical solution of the disclosure
Monotropic type, these simple variants belong to the protection scope of the disclosure.
It is further to note that specific technical features described in the above specific embodiments, in not lance
In the case where shield, it can be combined in any appropriate way.In order to avoid unnecessary repetition, the disclosure to it is various can
No further explanation will be given for the combination of energy.
In addition, any combination can also be carried out between a variety of different embodiments of the disclosure, as long as it is without prejudice to originally
Disclosed thought equally should be considered as disclosure disclosure of that.
Claims (13)
1. a kind of monitoring and managing method, which is characterized in that be applied to cloud server, which comprises
The suspicious monitoring information that monitoring client is sent is received, the suspicious monitoring information is that the monitoring client is based on single feature identification mould
Type identifies the obtained monitoring information with anomalous event from collected monitoring information, and the list feature identification model is benefit
The single features in monitoring information are used to obtain as training sample training；
The suspicious monitoring information is inputted into multiple features fusion model, obtains the recognition result of anomalous event, the multiple features melt
Molding type is using the anomalous event label of multiple features and the monitoring information in monitoring information as training sample to instruction
It gets.
2. the method according to claim 1, wherein the method also includes:
If the recognition result shows the monitoring client, there are anomalous events, and the suspicious monitoring information is sent to purpose
End, to indicate that the destination shows the suspicious monitoring information.
3. according to the method described in claim 2, it is characterized in that, the method also includes:
Receive the manual identified result to the suspicious monitoring information that the destination is sent；
Training sample pair is generated according to the suspicious monitoring information and the manual identified result；
Using the training sample to the update multiple features fusion model.
4. according to the method described in claim 2, it is characterized in that, the method also includes:
The manual identified result is sent to the monitoring client, the manual identified result is for described in monitoring client update
Single feature identification model.
5. method according to any one of claims 1 to 4, which is characterized in that the monitoring information include image information,
One or more of voice messaging, text information, location information and movement state information.
6. a kind of maintenance device, which is characterized in that be applied to cloud server, described device includes:
First receiving module is configured as receiving the suspicious monitoring information that monitoring client is sent, and the suspicious monitoring information is described
Monitoring client identifies the obtained monitoring information with anomalous event based on single feature identification model from collected monitoring information,
The list feature identification model is obtained using the single features in monitoring information as training sample training；
Identification module is configured as the suspicious monitoring information input multiple features fusion model obtaining the identification of anomalous event
As a result, the multiple features fusion model is the anomalous event label using multiple features and the monitoring information in monitoring information
Training is obtained as training sample.
7. device according to claim 6, which is characterized in that described device further include:
First sending module, be configured as the recognition result show the monitoring client there are when anomalous event, will described in can
Doubtful monitoring information is sent to destination, to indicate that the destination shows the suspicious monitoring information.
8. device according to claim 7, which is characterized in that described device further include:
Second receiving module is configured as receiving the manual identified knot to the suspicious monitoring information that the destination is sent
Fruit；
Generation module is configured as generating training sample pair according to the suspicious monitoring information and the manual identified result；
Update module is configured as using the training sample to the update multiple features fusion model.
9. device according to claim 7, which is characterized in that described device further include:
Second sending module is configured as the manual identified result being sent to the monitoring client, the manual identified result
Single feature identification model is updated for the monitoring client.
10. the device according to any one of claim 6~9, which is characterized in that the monitoring information includes image letter
One or more of breath, voice messaging, text information, location information and movement state information.
11. computer readable storage medium is stored thereon with computer program, which is characterized in that the program is executed by processor
The step of any one of Shi Shixian Claims 1 to 5 the method.
12. a kind of cloud server characterized by comprising
Memory is stored thereon with computer program；
Processor, for executing the computer program in the memory, to realize any one of Claims 1 to 5 institute
The step of stating method.
13. a kind of supervisory systems, which is characterized in that including cloud server described in monitoring client and claim 12.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810982884.3A CN109327328A (en) | 2018-08-27 | 2018-08-27 | Monitoring and managing method, device, system, cloud server and storage medium |
| PCT/CN2019/083942 WO2020042637A1 (en) | 2018-08-27 | 2019-04-23 | Supervision method, apparatus and system, cloud server and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810982884.3A CN109327328A (en) | 2018-08-27 | 2018-08-27 | Monitoring and managing method, device, system, cloud server and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109327328A true CN109327328A (en) | 2019-02-12 |
Family
ID=65264455
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810982884.3A Pending CN109327328A (en) | 2018-08-27 | 2018-08-27 | Monitoring and managing method, device, system, cloud server and storage medium |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN109327328A (en) |
| WO (1) | WO2020042637A1 (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110222846A (en) * | 2019-05-13 | 2019-09-10 | 中国科学院计算技术研究所 | A kind of the information safety protection method and information security system of Internet terminal |
| WO2020042637A1 (en) * | 2018-08-27 | 2020-03-05 | 深圳前海达闼云端智能科技有限公司 | Supervision method, apparatus and system, cloud server and storage medium |
| CN110866543A (en) * | 2019-10-18 | 2020-03-06 | 支付宝(杭州)信息技术有限公司 | Picture detection and picture classification model training method and device |
| CN111786858A (en) * | 2020-07-06 | 2020-10-16 | 三星（中国）半导体有限公司 | Method for diagnosing abnormal data, user terminal equipment and cloud server |
| CN111858973A (en) * | 2020-07-30 | 2020-10-30 | 北京达佳互联信息技术有限公司 | Multimedia event information detection method, device, server and storage medium |
| CN113112724A (en) * | 2021-03-19 | 2021-07-13 | 深圳市海吉星智慧城市科技有限公司 | Oil tank truck anti-theft method based on machine learning |
| CN113103946A (en) * | 2021-03-19 | 2021-07-13 | 深圳恩联科技有限公司 | Oil tank truck anti-theft system based on machine learning |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102169544A (en) * | 2011-04-18 | 2011-08-31 | 苏州市慧视通讯科技有限公司 | Face-shielding detecting method based on multi-feature fusion |
| CN102592146A (en) * | 2011-12-28 | 2012-07-18 | 浙江大学 | Face detection and camera tripod control method applied to video monitoring |
| CN103020990A (en) * | 2012-12-06 | 2013-04-03 | 华中科技大学 | Moving object detecting method based on graphics processing unit (GPU) |
| CN104159071A (en) * | 2014-07-11 | 2014-11-19 | 深圳瞭望通达科技有限公司 | Intelligent target identification device, system and method based on cloud service |
| US20150164377A1 (en) * | 2013-03-13 | 2015-06-18 | Vaidhi Nathan | System and method of body motion analytics recognition and alerting |
| CN204759452U (en) * | 2015-07-09 | 2015-11-11 | 华南理工大学 | Traffic conflagration detecting system based on many characteristics of video smog fuse |
| CN105205728A (en) * | 2015-08-26 | 2015-12-30 | 国家电网公司 | Audio and video integration detection method orienting electric power machine room monitoring |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9466009B2 (en) * | 2013-12-09 | 2016-10-11 | Nant Holdings Ip. Llc | Feature density object classification, systems and methods |
| CN106407984B (en) * | 2015-07-31 | 2020-09-11 | 腾讯科技（深圳）有限公司 | Target object identification method and device |
| CN108259830B (en) * | 2018-01-25 | 2019-08-23 | 深圳冠思大数据服务有限公司 | Mouse based on Cloud Server suffers from intelligent monitor system and method |
| CN109327328A (en) * | 2018-08-27 | 2019-02-12 | 深圳前海达闼云端智能科技有限公司 | Monitoring and managing method, device, system, cloud server and storage medium |
-
2018
- 2018-08-27 CN CN201810982884.3A patent/CN109327328A/en active Pending
-
2019
- 2019-04-23 WO PCT/CN2019/083942 patent/WO2020042637A1/en active Application Filing
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102169544A (en) * | 2011-04-18 | 2011-08-31 | 苏州市慧视通讯科技有限公司 | Face-shielding detecting method based on multi-feature fusion |
| CN102592146A (en) * | 2011-12-28 | 2012-07-18 | 浙江大学 | Face detection and camera tripod control method applied to video monitoring |
| CN103020990A (en) * | 2012-12-06 | 2013-04-03 | 华中科技大学 | Moving object detecting method based on graphics processing unit (GPU) |
| US20150164377A1 (en) * | 2013-03-13 | 2015-06-18 | Vaidhi Nathan | System and method of body motion analytics recognition and alerting |
| CN104159071A (en) * | 2014-07-11 | 2014-11-19 | 深圳瞭望通达科技有限公司 | Intelligent target identification device, system and method based on cloud service |
| CN204759452U (en) * | 2015-07-09 | 2015-11-11 | 华南理工大学 | Traffic conflagration detecting system based on many characteristics of video smog fuse |
| CN105205728A (en) * | 2015-08-26 | 2015-12-30 | 国家电网公司 | Audio and video integration detection method orienting electric power machine room monitoring |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020042637A1 (en) * | 2018-08-27 | 2020-03-05 | 深圳前海达闼云端智能科技有限公司 | Supervision method, apparatus and system, cloud server and storage medium |
| CN110222846A (en) * | 2019-05-13 | 2019-09-10 | 中国科学院计算技术研究所 | A kind of the information safety protection method and information security system of Internet terminal |
| CN110222846B (en) * | 2019-05-13 | 2021-07-20 | 中国科学院计算技术研究所 | Information security method and information security system for internet terminal |
| CN110866543A (en) * | 2019-10-18 | 2020-03-06 | 支付宝(杭州)信息技术有限公司 | Picture detection and picture classification model training method and device |
| CN110866543B (en) * | 2019-10-18 | 2022-07-15 | 支付宝(杭州)信息技术有限公司 | Picture detection and picture classification model training method and device |
| CN111786858A (en) * | 2020-07-06 | 2020-10-16 | 三星（中国）半导体有限公司 | Method for diagnosing abnormal data, user terminal equipment and cloud server |
| CN111786858B (en) * | 2020-07-06 | 2022-04-15 | 三星（中国）半导体有限公司 | Method for diagnosing abnormal data, user terminal equipment and cloud server |
| CN111858973A (en) * | 2020-07-30 | 2020-10-30 | 北京达佳互联信息技术有限公司 | Multimedia event information detection method, device, server and storage medium |
| CN113112724A (en) * | 2021-03-19 | 2021-07-13 | 深圳市海吉星智慧城市科技有限公司 | Oil tank truck anti-theft method based on machine learning |
| CN113103946A (en) * | 2021-03-19 | 2021-07-13 | 深圳恩联科技有限公司 | Oil tank truck anti-theft system based on machine learning |
| CN113103946B (en) * | 2021-03-19 | 2023-08-01 | 深圳恩联科技有限公司 | Tank truck anti-theft system based on machine learning |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2020042637A1 (en) | 2020-03-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109327328A (en) | Monitoring and managing method, device, system, cloud server and storage medium | |
| US11328163B2 (en) | Methods and apparatus for automated surveillance systems | |
| US10812761B2 (en) | Complex hardware-based system for video surveillance tracking | |
| US8368754B2 (en) | Video pattern recognition for automating emergency service incident awareness and response | |
| EP3750100B1 (en) | Device, system and method for generating an alert and an automatic search for a candidate subject | |
| CN210691345U (en) | Cloud intelligence safety inspection auxiliary system based on degree of depth learning | |
| KR102384299B1 (en) | Cctv camera device having assault detection function and method for detecting assault based on cctv image performed | |
| CN109309808A (en) | A kind of monitoring system and method based on recognition of face | |
| CN107888887A (en) | A kind of video monitoring method for early warning and system for monitoring gas pipeline damage from third-party | |
| US9622048B2 (en) | SNS based incident management | |
| US20210360406A1 (en) | Internet-of-things device classifier | |
| CN111294561A (en) | Video-based online patrol method, electronic device, storage medium and system | |
| KR20170084657A (en) | System and method for generating narrative report based on video recognition and event trancking | |
| CN109978215B (en) | Patrol management method and device | |
| KR20180118979A (en) | Method and apparatus for risk detection, prediction, and its correspondence for public safety based on multiple complex information | |
| CN106097220A (en) | A kind of prediction scheme sends method and device | |
| CN113269039A (en) | On-duty personnel behavior identification method and system | |
| KR101548729B1 (en) | Method for porviding safe return service | |
| KR102344435B1 (en) | Safety control service system unsing artifical intelligence | |
| KR102228933B1 (en) | Method for providing online to offline based crime prevention service offering private bodyguard for single family | |
| US20210350116A1 (en) | System and method for identifying persons-of-interest | |
| CN105978999A (en) | Alarm method and device | |
| KR20200059643A (en) | ATM security system based on image analyses and the method thereof | |
| CN112699328A (en) | Network point service data processing method, device, system, equipment and storage medium | |
| US20210200826A1 (en) | Appearance search for incident response |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190212 |