CN110517427A - A kind of warning message recognition methods, server and system - Google Patents
A kind of warning message recognition methods, server and system Download PDFInfo
- Publication number
- CN110517427A CN110517427A CN201810497474.XA CN201810497474A CN110517427A CN 110517427 A CN110517427 A CN 110517427A CN 201810497474 A CN201810497474 A CN 201810497474A CN 110517427 A CN110517427 A CN 110517427A
- Authority
- CN
- China
- Prior art keywords
- human detection
- humanoid
- alarm
- warning message
- picture
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B13/00—Burglar, theft or intruder alarms
- G08B13/18—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength
- G08B13/189—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength using passive radiation detection systems
- G08B13/194—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength using passive radiation detection systems using image scanning and comparing systems
- G08B13/196—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength using passive radiation detection systems using image scanning and comparing systems using television cameras
- G08B13/19602—Image analysis to detect motion of the intruder, e.g. by frame subtraction
Abstract
Warning message recognition methods provided by the present application is applied to server, which comprises when receiving the warning message based on mobile detection event of headend equipment transmission, obtains multiple alarm pictures that warning message is directed to；The alarm picture is headend equipment collected picture in away from mobile detection event triggering moment preset time；Human detection is carried out to every alarm picture using a variety of Human detection algorithms；According to the humanoid quantity that every alarm picture that every kind of Human detection algorithm identifies includes, the corresponding dispersion degree of every kind of Human detection algorithm is calculated；The humanoid quantity that every alarm picture that the foundation humanoid recognizer of dispersion degree the smallest first identifies includes, it is determined whether there are invaders；If so, determining that the warning message is effective warning message.Warning message recognition methods, server and system provided by the present application can screen the warning message that headend equipment reports, and reduce a possibility that server makes invalid operation.
Description
Technical field
This application involves field of video monitoring more particularly to a kind of warning message recognition methods, server and system.
Background technique
Video monitoring system is the important component of safety and protection system, is a kind of stronger comprehensive system of prevention ability
System.Currently, video monitoring system is intuitive with it, accurate, timely with the progress of social economy developed rapidly with science and technology
It is widely used in many occasions with abundant in content feature.However, the scale application of video monitoring causes to go to distinguish by people
It is perfectly safe not to be difficult to accomplish.
Existing Video Surveillance Alarm System, headend equipment usually carries out mobile detection to area to be monitored, and then is detecing
Mobile detection event is triggered when having measured mobile object, and reports alarm signal to server when detecting mobile detection event
Breath.In this way, server when receiving the warning message, is determined there are invader, alarm operation is executed.For example, to client
It sends for reminding reminder message of the current time there are invader, for another example linkage control headend equipment rotates, to be tracked into
The person of invading etc..
But due to weather, illumination, the factors such as shadow and chaotic interference, there are certain rate of false alarms for headend equipment.And it takes
Business device is performed both by alarm operation for each warning message.In this way, some alarm operations are actually the alarm signal for wrong report
The operation made is ceased, belongs to invalid operation, there are problems that the wasting of resources.Therefore, currently, there is an urgent need for provide a kind of alarm signal
Cease recognition methods.
Summary of the invention
In view of this, the application provides a kind of warning message recognition methods, server and system, to report to headend equipment
Warning message screened, reduce server a possibility that making invalid operation.
The application first aspect provides a kind of warning message recognition methods, and the method is applied to server, the method
Include:
When receiving the warning message based on mobile detection event of headend equipment transmission, the warning message needle is obtained
Pair multiple alarm pictures；The alarm picture is the headend equipment when default away from the mobile detection event triggering moment
Interior collected picture；
Human detection is carried out to every alarm picture using a variety of Human detection algorithms；
According to the humanoid quantity that every alarm picture that every kind of Human detection algorithm identifies includes, every kind of humanoid knowledge is calculated
The corresponding dispersion degree of other algorithm；
The humanoid quantity that every alarm picture that the foundation humanoid recognizer of dispersion degree the smallest first identifies includes,
Determine whether there is invader；
If so, determining that the warning message is effective warning message.
The application second aspect provides a kind of server, and the server includes: communication unit, memory and processor；
Wherein,
The communication unit, for receive headend equipment transmission the warning message based on mobile detection event when,
Obtain multiple alarm pictures that the warning message is directed to；Wherein, the alarm picture is the headend equipment away from the shifting
Collected picture in dynamic detecting event triggering moment preset time；
The memory, for storing the alarm picture；
The processor, is used for:
Human detection is carried out to every alarm picture using a variety of Human detection algorithms；
According to the humanoid quantity that every alarm picture that every kind of Human detection algorithm identifies includes, every kind of humanoid knowledge is calculated
The corresponding dispersion degree of other algorithm；
The humanoid quantity that every alarm picture that the foundation humanoid recognizer of dispersion degree the smallest first identifies includes,
Determine whether there is invader；
If so, determining that the warning message is effective warning message.
The application third aspect provides a kind of computer storage medium, is stored thereon with computer program, described program quilt
Processor realizes the step of any the method that the application first aspect provides when executing.
The application fourth aspect provides a kind of warning message identifying system, and the system comprises headend equipments and server；
Wherein,
The headend equipment, for when detecting mobile detection event, Xiang Suoshu server to be sent based on mobile detection
The warning message of event；
The server, is used for:
When receiving the warning message, multiple alarm pictures that the warning message is directed to are obtained；The alarm figure
Piece is the headend equipment collected picture in away from the mobile detection event triggering moment preset time；
Human detection is carried out to every alarm picture using a variety of Human detection algorithms；
According to the humanoid quantity that every alarm picture that every kind of Human detection algorithm identifies includes, every kind of humanoid knowledge is calculated
The corresponding dispersion degree of other algorithm；
The humanoid quantity that every alarm picture that the foundation humanoid recognizer of dispersion degree the smallest first identifies includes,
Determine whether there is invader；
If so, determining that the warning message is that effective alarm is normal.
Warning message recognition methods, server and system provided by the present application, server are receiving headend equipment transmission
The warning message based on mobile detection event when, by obtaining multiple alarm pictures for being directed to of warning message, and using a variety of
Human detection algorithm carries out Human detection, and then the every report identified according to every kind of Human detection algorithm to every alarm picture
The humanoid quantity that alert picture includes calculates the corresponding dispersion degree of every kind of Human detection algorithm, and the smallest according to dispersion degree
The humanoid quantity that every alarm picture that first humanoid recognizer identifies includes, it is determined whether there are invaders, thus
Determine that there are when invader, determine that warning message is effective warning message.This provides a kind of examination sides of warning message
Method can screen warning message, to reduce a possibility that server makes invalid operation, save resource.In addition, this implementation
Warning message recognition methods, server and the system that example provides carry out a variety of alarm pictures using a variety of Human detection algorithms
Human detection, and then the humanoid quantity identified according to the humanoid recognizer of the lesser target of dispersion, it is determined whether exist into
The person of invading can accurately screen warning message based on Human detection result in this way, Human detection result accuracy is higher.
Detailed description of the invention
Fig. 1 is the application scenarios signal of the warning message recognition methods provided by the present application shown in an exemplary embodiment
Figure；
Fig. 2 is the flow chart of warning message recognition methods embodiment one provided by the present application；
Fig. 3 is the flow chart of warning message recognition methods embodiment two provided by the present application；
Fig. 4 is the flow chart of warning message recognition methods embodiment three provided by the present application；
Fig. 5 is the interaction diagrams of the headend equipment and server shown in one exemplary embodiment of the application；
Fig. 6 is the structural schematic diagram of server example one provided by the present application.
Specific embodiment
Example embodiments are described in detail here, and the example is illustrated in the accompanying drawings.Following description is related to
When attached drawing, unless otherwise indicated, the same numbers in different drawings indicate the same or similar elements.Following exemplary embodiment
Described in embodiment do not represent all embodiments consistent with the application.On the contrary, they be only with it is such as appended
The example of the consistent device and method of some aspects be described in detail in claims, the application.
It is only to be not intended to be limiting the application merely for for the purpose of describing particular embodiments in term used in this application.
It is also intended in the application and the "an" of singular used in the attached claims, " described " and "the" including majority
Form, unless the context clearly indicates other meaning.It is also understood that term "and/or" used herein refers to and wraps
It may be combined containing one or more associated any or all of project listed.
It will be appreciated that though various information, but this may be described using term first, second, third, etc. in the application
A little information should not necessarily be limited by these terms.These terms are only used to for same type of information being distinguished from each other out.For example, not departing from
In the case where the application range, the first information can also be referred to as the second information, and similarly, the second information can also be referred to as
One information.Depending on context, word as used in this " if " can be construed to " ... when " or " when ...
When " or " in response to determination ".
The application provides a kind of warning message recognition methods, system and server, with the alarm signal reported to headend equipment
Breath is screened, and a possibility that server makes invalid operation is reduced.
Fig. 1 is the warning message recognition methods provided by the present application and systematic difference field shown in an exemplary embodiment
Scape.Please refer to Fig. 1, application scenario diagram shown in Fig. 1, monitoring system includes headend equipment 1, server 2 and client 3, existing
In technology, headend equipment 1 sends warning message after triggering mobile detection event, to server 2, and further, server 2 exists
When receiving the warning message, send to client 3 for reminding reminder message of the current time there are invader.In this way, In
When warning message is the warning message of wrong report, it is clear that server 2 will send invalid reminder message to client 3, cause resource
Waste.
This application provides a kind of warning message recognition methods, to be screened to the warning message that headend equipment reports,
In this way, server can just execute alarm operation, can reduce server and make when determining warning message is effective warning message
A possibility that invalid operation, saves resource.
Several specific embodiments are given below, for the technical solution of the application to be discussed in detail, below these are specific
Embodiment can be combined with each other, the same or similar concept or process may be repeated no more in certain embodiments.
Fig. 2 is the flow chart of warning message recognition methods embodiment one provided by the present application.Method provided in this embodiment,
Applied in server shown in FIG. 1.Referring to figure 2., method provided in this embodiment may include:
S201, receive headend equipment transmission the warning message based on mobile detection event when, obtain above-mentioned alarm
Multiple alarm pictures that information is directed to；Above-mentioned alarm picture is above-mentioned headend equipment away from above-mentioned mobile detection event triggering moment
Collected picture in preset time.
Specifically, headend equipment after detecting mobile detection event, can grab it is pre- away from mobile detection event triggering moment
If in the time (including before mobile detection event triggering moment in preset time and between mobile detection event triggering moment it
Afterwards in preset time) multiple alarm pictures.
It should be noted that after headend equipment grabs alarm picture, can be incited somebody to action in the possible implementation of the application one
Alarm picture is stored in image storage unit, which can be located locally, or is located in storage server, in turn
The storage address of alarm picture is carried in warning message, so that server is when receiving warning message, according to warning message
The storage address of middle carrying obtains alarm picture.Certainly, in another possible implementation of the application, headend equipment is being grabbed
After alarm picture, alarm picture directly can be sent to server.
S202, Human detection is carried out to every alarm picture using a variety of Human detection algorithms.
Specifically, Human detection algorithm is the Human detection algorithm based on deep learning.Below with " multiple alarm picture packets
Include 5 alarm pictures, a variety of Human detection algorithms include two kinds of Human detection algorithms " for be illustrated.In addition, for convenient for area
Point, this 5 alarm pictures are denoted as alarm picture 1, alarm picture 2, alarm picture 3, alarm picture 4 and alarm picture 5 respectively,
Both Human detection algorithms are denoted as Human detection algorithm A and Human detection algorithm B respectively.In this step, just it is respectively adopted
Human detection algorithm A and B carries out Human detection to each alarm picture.
S203, the humanoid quantity for including according to every alarm picture that every kind of Human detection algorithm identifies calculate every kind
The corresponding dispersion degree of Human detection algorithm.
Specifically, can be returned humanoid in picture after carrying out Human detection to alarm picture using Human detection algorithm
Rectangular coordinates.Therefore, every report that every kind of Human detection algorithm identifies can be determined according to the quantity of the rectangular coordinates of return
The humanoid quantity that alert picture includes.In conjunction with above example, for example, table 1 is every kind shown in one exemplary embodiment of the application
The humanoid quantity that every alarm picture that Human detection algorithm identifies includes, please refers to table 1, humanoid in the example shown in table 1
Recognizer A identifies that the humanoid quantity that alarm picture 1 includes is 0 ... ..., and Human detection algorithm B identifies that alarm picture 5 wraps
The humanoid quantity contained is 2.
The humanoid quantity that every alarm picture that the every kind of Human detection algorithm of table 1 identifies includes
Optionally, in the possible implementation of the application one, the specific implementation process of this step may include:
The humanoid quantity that the every alarm picture identified according to every kind of Human detection algorithm includes calculates every kind of humanoid knowledge
The mean square deviation of the corresponding humanoid quantity of other algorithm, and using above-mentioned mean square deviation as the corresponding discrete journey of every kind of Human detection algorithm
Degree.
Specifically, the mean square deviation of the corresponding humanoid quantity of every kind of Human detection algorithm, calculates according to following formula:
Wherein, σ is the mean square deviation of the corresponding humanoid quantity of humanoid recognizer；
N is the quantity of alarm picture；
xiThe humanoid quantity for including for i-th alarm picture that humanoid recognizer identifies；
μ is the average value for the humanoid quantity that each alarm picture that humanoid recognizer identifies includes.
For example, being computed in the example shown in table 1, the mean square deviation for obtaining the corresponding humanoid quantity of Human detection algorithm A is
The mean square deviation of the corresponding humanoid quantity of 4.242, Human detection algorithm B is 1.224.
It should be noted that in another possible implementation of the application, it, can also be humanoid according to every kind in this step
The humanoid quantity that every alarm picture that recognizer identifies includes calculates the corresponding humanoid quantity of every kind of Human detection algorithm
Variance, and using above-mentioned variance as the corresponding dispersion degree of every kind of Human detection algorithm.Specific formula in relation to calculating variance
It may refer to description in the prior art, details are not described herein again.
S204, the every alarm picture identified according to the humanoid recognizer of dispersion degree the smallest first include humanoid
Quantity, it is determined whether there are invaders.
Specifically, the example in conjunction with shown in table 1, in this step, every picture for just being identified according to Human detection algorithm B
The humanoid quantity for including, it is determined whether there are invaders.
Optionally, in the possible implementation of the application one, invader can be determined whether there is as follows, it should
Method may include:
When the humanoid quantity that any one alarm picture that the first humanoid recognizer identifies includes is greater than 0, determine
There are invaders；
When all alarm pictures that the first humanoid recognizer identifies include it is humanoid equal in number in 0 when, determine not
There are invaders.
In conjunction with above example, in this example, determine that there are invaders.
S205, if so, determining that above-mentioned warning message is effective warning message.
Further, when invader is not present in determination, at this point, determining that above-mentioned warning message is invalid alarm signal
Breath determines that above-mentioned warning message is the warning message of wrong report.
Method provided in this embodiment, server is in the alarm based on mobile detection event for receiving headend equipment transmission
When information, multiple alarm pictures being directed to by obtaining warning message, and every alarm is schemed using a variety of Human detection algorithms
Piece carries out Human detection, and then the humanoid quantity for including according to every alarm picture that every kind of Human detection algorithm identifies, meter
The corresponding dispersion degree of every kind of Human detection algorithm is calculated, and identified according to the humanoid recognizer of dispersion degree the smallest first
The humanoid quantity that every alarm picture includes, it is determined whether there are invaders, to determine alarm determining there are when invader
Information is effective warning message.This provides a kind of discriminating methods of warning message, can screen to warning message,
To reduce a possibility that server makes invalid operation, resource is saved.In addition, method provided in this embodiment, using a variety of people
Shape recognizer carries out Human detection to a variety of alarm pictures, and then according to the humanoid recognizer identification of the lesser target of dispersion
Humanoid quantity out, it is determined whether there are invaders, in this way, Human detection result accuracy is higher, can be based on Human detection knot
Fruit accurately screens warning message.
Fig. 3 is the flow chart of warning message recognition methods embodiment two provided by the present application.The executing subject of the present embodiment
For server shown in FIG. 1.Referring to figure 3., on the basis of the above embodiments, method provided in this embodiment, step S202
Before, can also include:
S301, according to above-mentioned alarm picture, obtain the scene information of above-mentioned headend equipment.
Specifically, scene information may include at least one in following information: environmental information, light information and scene class
Not.
For example, scene information may include environmental information in a possible implementation, which includes interior
Environment and outdoor environment.In this step, the environmental information of headend equipment is obtained using trained model in advance, for example, can
It is input in the preparatory trained model to alarm picture, obtains the environmental information of headend equipment.For example, in an embodiment,
The environmental information for obtaining headend equipment is indoor environment.
Certainly, in another possible implementation, scene information may include light information, which can wrap
Include visible light environment and infrared luminous environment.At this point, this step, can also be used the light that trained model in advance obtains headend equipment
Line information.
Further, in the application further possible implementation, scene information can be scene type, scene type
It may include supermarket's scene, campus scene, station scene and home scenarios etc..At this point, using preparatory trained scene class
The scene type of other identification model identification headend equipment.For example, in an embodiment, preparatory trained scene type identification model
Including 5 kinds of scene types, respectively supermarket's scene, campus scene, station scene, home scenarios and other scenes, at this point, can incite somebody to action
Alarm picture is input in the model, identifies the scene type of headend equipment, for example, identifying the scene type of headend equipment
For supermarket's scene.
It is illustrated so that scene information includes visible light environment and infrared luminous environment as an example below.For example, being obtained in this example
Scene information to headend equipment is visible light environment.
The corresponding relationship of scene information and preset scene information and Human detection algorithm that S302, basis are got,
Determine the humanoid recognizer of target corresponding with above-mentioned headend equipment.
Specifically, the corresponding relationship of scene information and Human detection algorithm is set according to actual needs, for example, can be according to
According to experience, the corresponding relationship of set scene information and Human detection algorithm, wherein Human detection corresponding with a scene information is calculated
Method is the higher Human detection algorithm of recognition accuracy in the scene information.
Table 2 is please referred to, for example, table 2 is corresponding with Human detection algorithm for the scene information shown in an exemplary embodiment
Relationship in the example shown in table 2, determines that the humanoid algorithm of target corresponding with headend equipment is humanoid knowledge in conjunction with above example
Other algorithm A, Human detection algorithm B and Human detection algorithm F.
The corresponding relationship of table 2 scene information and Human detection algorithm
Correspondingly, method provided in this embodiment, step S202, specifically includes: using the humanoid recognizer of above-mentioned target
Human detection is carried out to every alarm picture.
Method provided in this embodiment, using a variety of Human detection algorithms to every alarm picture carry out Human detection it
Before, by obtaining the scene information of headend equipment, and then according to the scene information and preset scene information that get and people
The corresponding relationship of shape recognizer determines the humanoid recognizer of target corresponding with above-mentioned headend equipment.In this way, using a variety of
When Human detection algorithm carries out Human detection to every alarm picture, every alarm is schemed using the humanoid recognizer of above-mentioned target
Piece carries out Human detection.In this way, being directed to the alarm picture of different headend equipments, carried out using different Human detection algorithms humanoid
The accuracy rate of Human detection can be improved in identification, improves the accuracy of warning message identification.
Fig. 4 is the flow chart of warning message recognition methods embodiment three provided by the present application.Method provided in this embodiment,
It can be applied in server shown in FIG. 1.Referring to figure 4., on the basis of the above embodiments, method provided in this embodiment,
Before step S202, can also include:
S401, according to above-mentioned alarm picture, obtain the scene information of above-mentioned headend equipment.
Specifically, the specific implementation process and realization principle of the step may refer to the description in preceding embodiment, herein
It repeats no more.
It include below scene type with scene information, scene type includes supermarket's scene, campus scene, station scene and family
It is illustrated for the scene of front yard.For example, the scene information for getting headend equipment is supermarket's scene in this example.
S402, it is calculated according to the scene information and preset scene information, Human detection algorithm and the Human detection that get
Corresponding relationship between the corresponding confidence level three of method, determine the humanoid recognizer of corresponding with above-mentioned headend equipment target and
The corresponding confidence level of the above-mentioned humanoid recognizer of target.
Specifically, the correspondence between scene information, Human detection algorithm and the corresponding confidence level three of Human detection algorithm
Relationship is set according to actual needs, in the present embodiment, is defined not to this.
Table 3 is please referred to, for example, table 3 is scene information, Human detection algorithm and the humanoid knowledge shown in an exemplary embodiment
Corresponding relationship between the corresponding confidence level three of other algorithm in the example shown in table 3, is determined with before in conjunction with above example
The humanoid algorithm of the corresponding target of end equipment is humanoid recognizer A and Human detection algorithm B, and Human detection algorithm A is corresponding
Confidence level is 80%, and the corresponding confidence level of Human detection algorithm B is 85%.
Corresponding relationship between 3 scene information of table, Human detection algorithm confidence level three corresponding with Human detection algorithm
Correspondingly, in the present embodiment, Human detection, packet are carried out to every alarm picture using a variety of Human detection algorithms
It includes:
Every alarm is schemed using the humanoid gender algorithm of above-mentioned target and the corresponding confidence level of the humanoid algorithm of above-mentioned target
Piece carries out Human detection.
Method provided in this embodiment, using a variety of Human detection algorithms to every alarm picture carry out Human detection it
Before, by obtaining the scene information of headend equipment, and then according to the scene information and preset scene information, humanoid got
Corresponding relationship between recognizer and the corresponding confidence level three of Human detection algorithm, determination are corresponding with above-mentioned headend equipment
The humanoid recognizer of target and the corresponding confidence level of above-mentioned Human detection algorithm.In this way, using a variety of Human detection algorithms
When carrying out Human detection to every alarm picture, using the humanoid recognizer of above-mentioned target and the humanoid recognizer pair of above-mentioned target
The confidence level answered carries out Human detection to every alarm picture.In this way, the accuracy rate of Human detection can be improved, warning message is improved
The accuracy of identification.
Optionally, in the possible implementation of the application one, after obtaining multiple alarm pictures that warning message is directed to,
The method also includes:
Above-mentioned alarm picture is pre-processed；Wherein, above-mentioned pretreatment includes decryption processing and/or dissection process.
Specifically, headend equipment after grabbing alarm picture, can encrypt alarm picture, for example, using non-right
Algorithm is claimed to encrypt alarm picture.Correspondingly, server is after getting encrypted alarm picture, just to the alarm figure
Piece is decrypted.
Further, headend equipment is also possible to for multiple alarm pictures to be integrated into a picture, at this point, server is obtaining
To after alarm picture, can to the alarm picture carry out dissection process (photo is split as plurality of pictures), with from this
Multiple alarm pictures are split out in picture.
It should be noted that in the possible implementation of the application one, receive headend equipment transmission based on shifting
When the warning message of dynamic detecting event, legitimate verification can also be carried out to the headend equipment, and then verifying the headend equipment
When legal, obtain multiple alarm pictures for being directed to of warning message, verify the headend equipment it is illegal when, ignore above-mentioned alarm signal
Breath.
Fig. 5 is the interaction diagrams of the headend equipment and server shown in one exemplary embodiment of the application.Below with reference to
Fig. 5 introduces warning message recognition methods provided by the present application.Referring to figure 5., warning message identification provided in this embodiment
Method may include:
S501, headend equipment send the report based on mobile detection event when detecting mobile detection event, to server
Alert information.
S502, above-mentioned server obtain multiple alarms that above-mentioned warning message is directed to when receiving above-mentioned warning message
Picture；Wherein, above-mentioned alarm picture is that above-mentioned headend equipment is adopted in away from above-mentioned mobile detection event triggering moment preset time
The picture collected.
S503, above-mentioned server carry out Human detection to every alarm picture using a variety of Human detection algorithms.
The humanoid number that S504, above-mentioned server include according to every alarm picture that every kind of Human detection algorithm identifies
Amount calculates the corresponding dispersion degree of every kind of Human detection algorithm.
Every alarm picture that S505, above-mentioned server are identified according to the humanoid recognizer of dispersion degree the smallest first
The humanoid quantity for including, it is determined whether there are invaders.
S506, above-mentioned server are determining that there are when invader, determine that above-mentioned warning message is that effective alarm is normal.
Specifically, specific implementation process and realization principle in relation to each step may refer to the description in preceding embodiment,
Details are not described herein again.
Method provided by the present application is described above.Server provided by the present application and system are retouched below
It states:
Fig. 6 is the structural schematic diagram of server example one provided by the present application.Fig. 6 is please referred to, it is provided in this embodiment
Server may include: communication unit 610, memory 620 and processor 630；Wherein,
The communication unit 610, in the warning message based on mobile detection event for receiving headend equipment transmission
When, obtain multiple alarm pictures that the warning message is directed to；Wherein, the alarm picture is the headend equipment away from described
Collected picture in mobile detection event triggering moment preset time；
The memory 620, for storing the alarm picture；
The processor 630, is used for:
Human detection is carried out to every alarm picture using a variety of Human detection algorithms；
According to the humanoid quantity that every alarm picture that every kind of Human detection algorithm identifies includes, every kind of humanoid knowledge is calculated
The corresponding dispersion degree of other algorithm；
The humanoid quantity that every alarm picture that the foundation humanoid recognizer of dispersion degree the smallest first identifies includes,
Determine whether there is invader；
If so, determining that the warning message is effective warning message.
The server of the present embodiment can be used for executing the technical solution of embodiment of the method shown in Fig. 2, realization principle and skill
Art effect is similar, and details are not described herein again.
Further, the processor 630 is also used to carry out every alarm picture using a variety of Human detection algorithms
Before Human detection, according to the alarm picture, the scene information of the headend equipment is obtained；And believed according to the scene got
The corresponding relationship of breath and preset scene information and Human detection algorithm determines that target corresponding with the headend equipment is humanoid
Recognizer；
It is described that Human detection is carried out to every alarm picture using a variety of Human detection algorithms, comprising:
Human detection is carried out to every alarm picture using the humanoid recognizer of the target.
Further, the processor 630 is also used to carry out every alarm picture using a variety of Human detection algorithms
Before Human detection, according to the alarm picture, the scene information of the headend equipment is obtained；And believed according to the scene got
Corresponding pass between breath and preset scene information, Human detection algorithm and the corresponding confidence level three of Human detection algorithm
System, the humanoid recognizer of determination target corresponding with the headend equipment and the corresponding confidence of the humanoid recognizer of the target
Degree；
It is described to carry out Human detection to every alarm picture using a variety of Human detection algorithms, comprising:
Every alarm is schemed using the humanoid gender algorithm of the target and the corresponding confidence level of the humanoid algorithm of the target
Piece carries out Human detection.
Further, the processor 630, specifically for the every alarm figure identified according to every kind of Human detection algorithm
The humanoid quantity that piece includes, calculates the mean square deviation of the corresponding humanoid quantity of every kind of Human detection algorithm, and the mean square deviation is made
For the corresponding dispersion degree of every kind of Human detection algorithm.
The processor 630, specifically for any one alarm picture identified when the described first humanoid recognizer
When the humanoid quantity for including is greater than 0, determine that there are invaders；When all alarm figures that the described first humanoid recognizer identifies
Piece includes humanoid equal in number when 0 to determine and invader is not present.
Further, the processor 630 is also used to pre-process the alarm picture；Wherein, the pretreatment
Including decryption processing and/or dissection process.
The application also provides a kind of computer storage medium, is stored thereon with computer program, described program is by processor
The step of any warning message recognition methods provided by the present application is realized when execution.
Specifically, being suitable for storing computer program instructions and the computer-readable medium of data including the non-of form of ownership
Volatile memory, medium and memory devices, for example including semiconductor memory devices (such as EPROM, EEPROM and flash memory
Equipment), disk (such as internal hard drive or removable disk), magneto-optic disk and CD ROM and DVD-ROM disk.
Please continue to refer to Fig. 1, the application also provides a kind of warning message identifying system, and the system comprises headend equipments 1
With server 2；Wherein,
The headend equipment 1, for when detecting mobile detection event, Xiang Suoshu server to be sent based on mobile detection
The warning message of event；
The server 2, is used for:
When receiving the warning message, multiple alarm pictures that the warning message is directed to are obtained；The alarm figure
Piece is the headend equipment collected picture in away from the mobile detection event triggering moment preset time；
Human detection is carried out to every alarm picture using a variety of Human detection algorithms；
According to the humanoid quantity that every alarm picture that every kind of Human detection algorithm identifies includes, every kind of humanoid knowledge is calculated
The corresponding dispersion degree of other algorithm；
The humanoid quantity that every alarm picture that the foundation humanoid recognizer of dispersion degree the smallest first identifies includes,
Determine whether there is invader；
If so, determining that the warning message is that effective alarm is normal.
Specifically, specific implementation process and realization principle in relation to each step may refer to the description in preceding embodiment,
Details are not described herein again.
Further, the server 2 is also used to carrying out people to every alarm picture using a variety of Human detection algorithms
Before shape identification, according to the alarm picture, the scene information of the headend equipment is obtained；And according to the scene information got
And the corresponding relationship of preset scene information and Human detection algorithm, determine the humanoid knowledge of target corresponding with the headend equipment
Other algorithm；
It is described that Human detection is carried out to every alarm picture using a variety of Human detection algorithms, comprising:
Human detection is carried out to every alarm picture using the humanoid recognizer of the target.
Further, the processor 2 is also used to carrying out people to every alarm picture using a variety of Human detection algorithms
Before shape identification, according to the alarm picture, the scene information of the headend equipment is obtained；And according to the scene information got
And the corresponding relationship between preset scene information, Human detection algorithm and the corresponding confidence level three of Human detection algorithm,
The humanoid recognizer of determination target corresponding with the headend equipment and the corresponding confidence level of the humanoid recognizer of the target；
It is described to carry out Human detection to every alarm picture using a variety of Human detection algorithms, comprising:
Every alarm is schemed using the humanoid gender algorithm of the target and the corresponding confidence level of the humanoid algorithm of the target
Piece carries out Human detection.
Further, the processor 2, specifically for the every alarm picture identified according to every kind of Human detection algorithm
The humanoid quantity for including, calculates the mean square deviation of the corresponding humanoid quantity of every kind of Human detection algorithm, and using the mean square deviation as
The corresponding dispersion degree of every kind of Human detection algorithm.
Further, the processor 2, specifically for any one report identified when the described first humanoid recognizer
When the humanoid quantity that alert picture includes is greater than 0, determine that there are invaders；When what the described first humanoid recognizer identified owns
Alarm picture includes humanoid equal in number when 0 to determine and invader is not present.
Further, the processor 2 is also used to after obtaining multiple alarm pictures that the warning message is directed to,
The alarm picture is pre-processed；Wherein, the pretreatment includes decryption processing and/or dissection process.
The foregoing is merely the preferred embodiments of the application, not to limit the application, all essences in the application
Within mind and principle, any modification, equivalent substitution, improvement and etc. done be should be included within the scope of the application protection.
Claims (10)
1. a kind of warning message recognition methods, which is characterized in that the method is applied to server, which comprises
When receiving the warning message based on mobile detection event of headend equipment transmission, obtain what the warning message was directed to
Multiple alarm pictures；The alarm picture is the headend equipment in away from the mobile detection event triggering moment preset time
Collected picture；
Human detection is carried out to every alarm picture using a variety of Human detection algorithms；
According to the humanoid quantity that every alarm picture that every kind of Human detection algorithm identifies includes, calculates every kind of Human detection and calculate
The corresponding dispersion degree of method；
The humanoid quantity that every alarm picture that the foundation humanoid recognizer of dispersion degree the smallest first identifies includes, determines
With the presence or absence of invader；
If so, determining that the warning message is effective warning message.
2. the method according to claim 1, wherein using a variety of Human detection algorithms to every alarm picture into
Before row Human detection, the method also includes:
According to the alarm picture, the scene information of the headend equipment is obtained；
According to the corresponding relationship of the scene information and preset scene information and Human detection algorithm that get, it is determining with it is described
The humanoid recognizer of the corresponding target of headend equipment；
It is described that Human detection is carried out to every alarm picture using a variety of Human detection algorithms, comprising:
Human detection is carried out to every alarm picture using the humanoid recognizer of the target.
3. the method according to claim 1, wherein described scheme every alarm using a variety of Human detection algorithms
Before piece carries out Human detection, the method also includes:
According to the alarm picture, the scene information of the headend equipment is obtained；
According to scene information and preset scene information, the Human detection algorithm and Human detection algorithm is corresponding sets got
Corresponding relationship between reliability three determines the humanoid recognizer of target corresponding with the headend equipment and the target person
The corresponding confidence level of shape recognizer；
It is described to carry out Human detection to every alarm picture using a variety of Human detection algorithms, comprising:
Using the humanoid gender algorithm of the target and the corresponding confidence level of the humanoid algorithm of the target to every alarm picture into
Row Human detection.
4. the method according to claim 1, wherein every identified according to every kind of Human detection algorithm
The humanoid quantity that alarm picture includes calculates the corresponding dispersion degree of every kind of Human detection algorithm, comprising:
The humanoid quantity that the every alarm picture identified according to every kind of Human detection algorithm includes calculates every kind of Human detection and calculates
The mean square deviation of the corresponding humanoid quantity of method, and using the mean square deviation as the corresponding dispersion degree of every kind of Human detection algorithm.
5. the method according to claim 1, wherein described calculate according to lesser first Human detection of dispersion degree
The humanoid quantity that every alarm picture that method identifies includes, it is determined whether there are invaders, comprising:
When the humanoid quantity that any one alarm picture that the described first humanoid recognizer identifies includes is greater than 0, determine
There are invaders；
When all alarm pictures that the described first humanoid recognizer identifies include it is humanoid equal in number in 0 when, determine not
There are invaders.
6. the method according to claim 1, wherein described multiple alarm figures for obtaining the warning message and being directed to
After piece, the method also includes:
The alarm picture is pre-processed；Wherein, the pretreatment includes decryption processing and/or dissection process.
7. a kind of server, which is characterized in that the server includes: communication unit, memory and processor；Wherein,
The communication unit, for obtaining when receiving the warning message based on mobile detection event of headend equipment transmission
Multiple alarm pictures that the warning message is directed to；Wherein, the alarm picture is that the headend equipment is detectd away from the movement
Collected picture in survey event triggering moment preset time；
The memory, for storing the alarm picture；
The processor, is used for:
Human detection is carried out to every alarm picture using a variety of Human detection algorithms；
According to the humanoid quantity that every alarm picture that every kind of Human detection algorithm identifies includes, calculates every kind of Human detection and calculate
The corresponding dispersion degree of method；
The humanoid quantity that every alarm picture that the foundation humanoid recognizer of dispersion degree the smallest first identifies includes, determines
With the presence or absence of invader；
If so, determining that the warning message is effective warning message.
8. server according to claim 7, which is characterized in that the processor is also used to using a variety of humanoid knowledges
Before other algorithm carries out Human detection to every alarm picture, according to the alarm picture, the scene of the headend equipment is obtained
Information；And according to the corresponding relationship of the scene information and preset scene information and Human detection algorithm that get, determine with
The corresponding humanoid recognizer of target of the headend equipment；
It is described that Human detection is carried out to every alarm picture using a variety of Human detection algorithms, comprising:
Human detection is carried out to every alarm picture using the humanoid recognizer of the target.
9. a kind of computer storage medium, is stored thereon with computer program, which is characterized in that described program is executed by processor
The step of any one of Shi Shixian claim 1-6 the method.
10. a kind of warning message identifying system, which is characterized in that the system comprises headend equipments and server；Wherein,
The headend equipment, for when detecting mobile detection event, Xiang Suoshu server to be sent based on mobile detection event
Warning message；
The server, is used for:
When receiving the warning message, multiple alarm pictures that the warning message is directed to are obtained；The alarm picture is
The headend equipment collected picture in away from the mobile detection event triggering moment preset time；
Human detection is carried out to every alarm picture using a variety of Human detection algorithms；
According to the humanoid quantity that every alarm picture that every kind of Human detection algorithm identifies includes, calculates every kind of Human detection and calculate
The corresponding dispersion degree of method；
The humanoid quantity that every alarm picture that the foundation humanoid recognizer of dispersion degree the smallest first identifies includes, determines
With the presence or absence of invader；
If so, determining that the warning message is effective warning message.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810497474.XA CN110517427B (en) | 2018-05-22 | 2018-05-22 | Alarm information identification method, server and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810497474.XA CN110517427B (en) | 2018-05-22 | 2018-05-22 | Alarm information identification method, server and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110517427A true CN110517427A (en) | 2019-11-29 |
| CN110517427B CN110517427B (en) | 2022-03-25 |
Family
ID=68622102
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810497474.XA Active CN110517427B (en) | 2018-05-22 | 2018-05-22 | Alarm information identification method, server and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110517427B (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080192980A1 (en) * | 2007-02-14 | 2008-08-14 | Samsung Electronics Co., Ltd. | Liveness detection method and apparatus of video image |
| CN101814222A (en) * | 2010-04-23 | 2010-08-25 | 泉州市科立信安防电子有限公司 | Intrusion detection device |
| CN104038375A (en) * | 2014-06-30 | 2014-09-10 | 成都广达电子股份有限公司 | Alarm processing analysis system and method of broadcasting and TV access network |
| CN104658152A (en) * | 2015-02-15 | 2015-05-27 | 西安交通大学 | Video-based moving object intrusion alarm method |
| CN105243773A (en) * | 2015-09-25 | 2016-01-13 | 国网山东省电力公司经济技术研究院 | Portable intelligent alarm fence and human proximity detection method |
| CN105830009A (en) * | 2013-12-20 | 2016-08-03 | 高通股份有限公司 | Selection and tracking of objects for display partitioning and clustering of video frames |
| CN106603999A (en) * | 2017-02-17 | 2017-04-26 | 上海创米科技有限公司 | Video monitoring alarming method and system |
| CN106647653A (en) * | 2016-11-23 | 2017-05-10 | 宁波市镇海百硕机械科技有限公司 | Real-time data acquisition system for textile workshop |
| CN106951867A (en) * | 2017-03-22 | 2017-07-14 | 成都擎天树科技有限公司 | Face identification method, device, system and equipment based on convolutional neural networks |
| CN108021898A (en) * | 2017-12-14 | 2018-05-11 | 西北工业大学 | A kind of personage's recognition methods again that study is measured based on adapted distance |
-
2018
- 2018-05-22 CN CN201810497474.XA patent/CN110517427B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080192980A1 (en) * | 2007-02-14 | 2008-08-14 | Samsung Electronics Co., Ltd. | Liveness detection method and apparatus of video image |
| CN101814222A (en) * | 2010-04-23 | 2010-08-25 | 泉州市科立信安防电子有限公司 | Intrusion detection device |
| CN105830009A (en) * | 2013-12-20 | 2016-08-03 | 高通股份有限公司 | Selection and tracking of objects for display partitioning and clustering of video frames |
| CN104038375A (en) * | 2014-06-30 | 2014-09-10 | 成都广达电子股份有限公司 | Alarm processing analysis system and method of broadcasting and TV access network |
| CN104658152A (en) * | 2015-02-15 | 2015-05-27 | 西安交通大学 | Video-based moving object intrusion alarm method |
| CN105243773A (en) * | 2015-09-25 | 2016-01-13 | 国网山东省电力公司经济技术研究院 | Portable intelligent alarm fence and human proximity detection method |
| CN106647653A (en) * | 2016-11-23 | 2017-05-10 | 宁波市镇海百硕机械科技有限公司 | Real-time data acquisition system for textile workshop |
| CN106603999A (en) * | 2017-02-17 | 2017-04-26 | 上海创米科技有限公司 | Video monitoring alarming method and system |
| CN106951867A (en) * | 2017-03-22 | 2017-07-14 | 成都擎天树科技有限公司 | Face identification method, device, system and equipment based on convolutional neural networks |
| CN108021898A (en) * | 2017-12-14 | 2018-05-11 | 西北工业大学 | A kind of personage's recognition methods again that study is measured based on adapted distance |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110517427B (en) | 2022-03-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Cucchiara et al. | A multi‐camera vision system for fall detection and alarm generation | |
| US9286778B2 (en) | Method and system for security system tampering detection | |
| JP2019216424A (en) | Surveillance system and surveillance method | |
| AU2019295856B2 (en) | Object tracking using disparate monitoring systems | |
| CN108073890A (en) | Action recognition in video sequence | |
| US9053621B2 (en) | Image surveillance system and image surveillance method | |
| US20070252693A1 (en) | System and method for surveilling a scene | |
| US10419728B2 (en) | Monitoring system having personal information protection function and method thereof | |
| CN107122743B (en) | Security monitoring method and device and electronic equipment | |
| GB2575282A (en) | Event entity monitoring network and method | |
| CN115004269B (en) | Monitoring device, monitoring method, and program | |
| CN107517391B (en) | Method and equipment for identifying abnormal live broadcast information in video live broadcast | |
| KR102264275B1 (en) | Violent behavior management system and method | |
| US11184476B2 (en) | Preventing photo image related risks | |
| TWI711319B (en) | Block chain-based smart alarm method and device, and electronic equipment | |
| CN116170566A (en) | Intelligent building monitoring management method and device, electronic equipment and storage medium | |
| US10880735B2 (en) | Method for detecting the possible taking of screenshots | |
| CN110517427A (en) | A kind of warning message recognition methods, server and system | |
| Gupta et al. | Multimodal complex event detection framework for wide area surveillance | |
| CN112949359A (en) | Convolutional neural network-based abnormal behavior identification method and device | |
| US20160086480A1 (en) | Crime evidence provider cum help seeker | |
| Hang et al. | Design and implementation of intelligent fire notification service using IP camera in smart home | |
| CN111696290A (en) | Security decision method and device, computing equipment and computer storage medium | |
| US11670080B2 (en) | Techniques for enhancing awareness of personnel | |
| KR20230066891A (en) | Smart video management device and method with situation awareness using correlation analysis between objects |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |