CN110378200A - A kind of intelligent security guard prompt apparatus and method for of Behavior-based control feature clustering - Google Patents
A kind of intelligent security guard prompt apparatus and method for of Behavior-based control feature clustering Download PDFInfo
- Publication number
- CN110378200A CN110378200A CN201910478050.3A CN201910478050A CN110378200A CN 110378200 A CN110378200 A CN 110378200A CN 201910478050 A CN201910478050 A CN 201910478050A CN 110378200 A CN110378200 A CN 110378200A
- Authority
- CN
- China
- Prior art keywords
- target
- frame
- boundary rectangle
- video
- prompt
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The present invention provides a kind of intelligent security guards of Behavior-based control feature clustering to prompt apparatus and method for, which comprises detects to each target in each frame video pictures, and marks each target with boundary rectangle；Record the parameter information of each target boundary rectangle in each frame video pictures；One section of video frame is intercepted, the changing value of the parameter information of the target boundary rectangle of the target boundary rectangle of each frame and the same target of former frame is calculated separately, forms a parameter variation value array for each target in this section of video frame；Obtained multiple arrays are executed into cluster calculation, class cluster is divided to the target in this section of video frame, choose the class cluster that number of targets is less than threshold value；The place frame for determining the target in selected class cluster provides security protection prompt.Security-protecting and monitoring method of the invention can not only reduce system operations amount, reduce hardware cost, can also further increase detection accuracy, guarantee the accuracy of detection abnormal behaviour.
Description
Technical field
The invention belongs to technical field of video monitoring, and in particular to a kind of intelligent security guard prompt of Behavior-based control feature clustering
Apparatus and method for.
Background technique
In smart city, by being widely distributed in the camera of city everywhere, video security monitoring is carried out towards crowd, is
One very important application.Analysis by backstage to security protection video picture therefrom can endanger public peace in discovery crowd
Complete and traffic order behavior, for example, fight, in violation of rules and regulations passage, it is crowded trample, and then give timely warning note,
So that front end sends rapidly administrative staff or police strength to be disposed, avoids greatly influencing and and endanger.
However, a cameras up to a hundred, each camera may be disposed for key areas such as square, transport hubs at present
The more than frame security protection video picture of real-time grasp shoot 20, sent backstage and was analyzed each second, thousands of frames per second even frames up to ten thousand
Picture, arithmetic speed require to reach millions of secondary per second, this brings great load to the analysis on backstage and warning note.Cause
This, it is desirable to reduce the operand of security protection video picture analysis, as far as possible to reduce load and hardware cost；It is in the prior art poly-
The acquisition of class algorithm, parameter information is complex, not abundant enough, is not able to satisfy the requirement to video analysis accuracy.
Summary of the invention
In view of this, the present invention provides a kind of apparatus and method for of the intelligent security guard of Behavior-based control feature clustering prompt,
Main purpose is to improve the accuracy of unusual checking, reduces system operations amount, reduce hardware cost.
To achieve the goals above, this application provides following technical schemes:
This application provides a kind of methods of the intelligent security guard of Behavior-based control feature clustering prompt, which comprises
Each target in each frame video pictures is detected, and marks each target with boundary rectangle；
Record the parameter information of each target boundary rectangle in each frame video pictures；
One section of video frame is intercepted, outside the target for calculating separately the target boundary rectangle of each frame and the same target of former frame
The changing value of the parameter information of rectangle is connect, forms a parameter variation value array for each target in this section of video frame；
The parameter variation value array of the target complete obtained from this section of video frame is executed into cluster calculation, to this section of video frame
In target divide class cluster, choose number of targets be less than threshold value class cluster；
The place frame for determining the target in selected class cluster provides security protection prompt.
Present invention also provides a kind of equipment of the intelligent security guard of Behavior-based control feature clustering prompt, the equipment includes:
Object detection unit: detecting each target in each frame video pictures, and is marked respectively with boundary rectangle
A target；
Target component determination unit: the parameter information of each target boundary rectangle in each frame video pictures is recorded；
Computing unit: one section of video frame of interception calculates separately the target boundary rectangle of each frame and the same mesh of former frame
The changing value of the parameter information of target target boundary rectangle forms a parameter variation value for each target in this section of video frame
Array；
Cluster cell: executing cluster calculation for the parameter variation value array of the target complete obtained from this section of video frame, right
Target in this section of video frame divides class cluster, chooses the class cluster that number of targets is less than threshold value；
Prompt unit: determining the place frame of the target in selected class cluster, provides security protection prompt.
It can be seen that the apparatus and method for that a kind of intelligent security guard of Behavior-based control feature clustering provided by the present application prompts,
It is the parameter information for marking target by a boundary rectangle, and recording boundary rectangle, by the target boundary rectangle phase of each frame
The changing value of the target boundary rectangle parameter information of target same for former frame, as the behavioural characteristic data of clustering,
Compared with prior art, the present invention is only acquired the parameter information of boundary rectangle, and parameter acquisition modes are simple and convenient, can
To reduce operand；Further, since boundary rectangle parameter information is abundant, by the target boundary rectangle of each frame relative to former frame
The changing value of the target boundary rectangle parameter information of same target, is analyzed as the behavioural characteristic data of clustering,
It can be improved the accuracy of clustering, improve result precision.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The embodiment of invention for those of ordinary skill in the art without creative efforts, can also basis
The attached drawing of offer obtains other attached drawings.
Fig. 1 is a kind of intelligence peace of Behavior-based control feature clustering of Behavior-based control feature clustering provided in an embodiment of the present invention
The flow diagram of anti-reminding method；
Fig. 2 is that a kind of process tested and analyzed in video to single target provided in an embodiment of the present invention is illustrated
Figure；
Fig. 3 is the process signal that a kind of pair of human target abnormal behaviour provided in an embodiment of the present invention carries out security protection prompt
Figure；
Fig. 4 is that a kind of intelligent security guard of Behavior-based control feature clustering provided in an embodiment of the present invention prompts the structure of equipment to show
It is intended to；
Fig. 5 is the structural schematic diagram of memory and processor in a kind of computer equipment provided in an embodiment of the present invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
As shown in Figure 1, being a kind of intelligence of the Behavior-based control feature clustering of Behavior-based control feature clustering provided by the invention
The flow diagram of security protection reminding method, which comprises
101: each target in each frame video pictures being detected, and is marked with boundary rectangle and tracks each mesh
Mark.
Each frame video pictures may have multiple targets, detect to multiple targets in video pictures, and will
The multiple targets detected are marked with boundary rectangle, each boundary rectangle just represents a target, in the present embodiment
Target can be personage, vehicle etc., and those skilled in the art can be configured according to specific requirements, here with no restrictions.
102: recording the parameter information of each target boundary rectangle in each frame video pictures.
It can be external as target using the transverse and longitudinal coordinate of the transverse and longitudinal coordinate on any vertex of target boundary rectangle or central point
The parameter information of rectangle, further, the parameter information of target boundary rectangle can also include that the width of target boundary rectangle is believed
Breath and elevation information.If there are multiple target boundary rectangles for present frame, to each target boundary rectangle recording parameters information.
103: one section of video frame of interception calculates separately the mesh of the target boundary rectangle of each frame and the same target of former frame
The changing value of each parameter information of boundary rectangle is marked, forms a parameter variation value array for each target in this section of video frame
As behavioural characteristic.
It is specifically as follows, intercepts one section of video frame, such as can intercepts 20 seconds, 30 seconds, one minute etc.；According to step 102
The parameter information of the target boundary rectangle of each target of each frame of middle record, by each frame in this section of video (in addition to mesh
Mark other than first existing video frame) each target boundary rectangle parameter information same mesh corresponding with its former frame
The parameter information of target target boundary rectangle compares, the change of the parameter information of each target boundary rectangle of available each frame
Change value, specifically, the changing value can be by the parameter information of the target boundary rectangle of present frame and the same target of former frame
The parameter information of target boundary rectangle do difference operation.In this section of video frame, same target may persistently will appear in multiple
In frame, thus the target boundary rectangle parameter information of same target may to be multiple, the changing value of parameter may also be it is multiple,
Therefore a multi-Dimensional parameters changing value array, the behavior as the target can be formed for each of this section of video frame target
Characteristic information.
For example, parameter information includes the cross of central point for the target boundary rectangle of one of target in the i-th frame
Coordinate Xi, ordinate Yi, boundary rectangle width Wi, boundary rectangle height Hi, by the mesh of itself and the same target in the (i-1)-th frame
The corresponding parameter information for marking boundary rectangle compares, and obtains the changing value Δ X of parameter informationi, Δ Yi, Δ Wi, Δ Hi, can incite somebody to action
It forms a four-dimensional array (Δ Xi, Δ Yi, Δ Wi, Δ Hi), as the behavioural characteristic of the target, for this section of video
Each of target, one group of 4 D data can be obtained.
Further, video frame can be intercepted according to certain frequency in advance, interception frequency can be less than video
The frequency of acquisition, such as video pictures can be intercepted according to the frequency of 20 frame per second, since interception frequency is lower, obtain
Video frame quantity be not it is very huge, can reduce the treating capacity of data, improve data-handling efficiency.
104: the parameter variation value array of the target complete obtained from this section of video frame being executed into cluster calculation, which is regarded
Target in frequency frame divides class cluster, chooses the class cluster that number of targets is less than threshold value.
For example, the multi-Dimensional parameters changing value array of the target complete in this section of video frame can be executed cluster calculation, draw
It is divided into N number of cluster, the similar target of behavioural characteristic can be generally gathered in the more class cluster of one or several targets, for behavior
Abnormal target, could be separately formed class cluster, therefore the class cluster less for destination number, the target for being included is very likely
There are abnormal behaviours.The less class cluster of destination number can be chosen to be analyzed, specifically, a threshold value can be set, work as class
When destination number is less than the threshold value in cluster, the video frame where the target in such cluster is extracted.
105: determining the place frame of the target in selected class cluster, provide security protection prompt.
For independently forming the target of class cluster, it is most likely that there are abnormal behaviours, it is therefore desirable to extract these targets place
Video frame checked, be specifically as follows determine these targets where frame number, according to frame number extract video frame, provide security protection
Prompt, so as to backstage, personnel check in time, detect whether to exist abnormal.
In the present invention, be less than video acquisition frequency due to intercepting the frequency of video to be detected, obtained number of pictures compared with
It is few, the operation pressure of system can be effectively reduced；The present invention is only acquired the parameter information of boundary rectangle, and parameter obtains
Mode is simple and convenient, can reduce operand；Meanwhile the present invention using the coordinate information of target boundary rectangle as parameter in addition to believing
Breath, also using the width information of target boundary rectangle and elevation information as parameter information, therefore when carrying out cluster calculation, has
The accuracy rate of cluster calculation can be improved in the characteristic information of very abundant.
Can have multiple targets in usual each video pictures, in specific implementation, often to each target require into
Row analysis, further, as the refinement and extension to above-described embodiment, Fig. 2 shows one kind provided in an embodiment of the present invention
The flow diagram that single target is tested and analyzed in video.Mainly include following procedure:
201: target is detected.Detection algorithm can be any detection algorithm in the prior art, such as may include
Background subtraction, edge detection, template matching and etc., here with no restrictions.
202: judging in video pictures with the presence or absence of target.Since what is monitored in video monitoring is a certain range of one group
Continuous picture, each target in picture are constantly moved, may at a time be appeared in picture, it is also possible to
The a certain moment leaves picture, it is therefore desirable to carry out real-time monitoring to each target, judge whether the target in picture exists.
203: if target exists, marking target with boundary rectangle, and continue step 204.Otherwise terminate to the target
Detection.
204: recording the parameter information of the boundary rectangle of each frame target.Parameter information may include target boundary rectangle
Coordinate information, width information, elevation information in the video frame etc., wherein coordinate information can be expressed as target boundary rectangle times
The coordinate information of one vertex or central point.Specifically, parameter information may include the horizontal seat of target boundary rectangle in video frame
Mark one or more of information, ordinate information, boundary rectangle width information, boundary rectangle elevation information；Wherein, selection
Parameter information is abundanter, and computational accuracy is higher, those skilled in the art can according to need in specific implementation to parameter into
Row selection.
205: the video frame in one section of monitor video picture of interception.Specifically, video can be intercepted according to certain frequency
Frame, such as according to the standard of 20 frame per second, the picture of interception one minute then includes 1200 frame safety monitoring video pictures, due to
Interception frequency is lower, and obtained number of pictures is less, it is possible to reduce the operation pressure of system.
206: calculating the changing value of the parameter information of the boundary rectangle of the target of each frame and former frame.Specifically, may be used
The corresponding parameter information of the boundary rectangle of each frame with the former frame target is done difference operation, because each frame can obtain
The changing value of the boundary rectangle parameter information of the target, therefore for video frame section existing for the target, available one group should
The parameter variation value array of target, using this array as the behavioural characteristic of target.
207: the parameter variation value array of the target is executed to cluster meter together with the parameter variation value array of other targets
It calculates.
For example, for the target, it is assumed that in the video of interception there are the frame number of the target be m, in addition to the target occur
Outside first frame, available m-1 group Parameters variation Value Data forms the parameter variation value array of the target；Similarly, for other
Target, it is also available to arrive corresponding parameter variation value array, whole arrays is executed into clustering, obtains calculated result.
208: judging that the target whether there is in the class cluster that number of targets is less than threshold value.If so, executing 209.Otherwise terminate
Detection.
Specifically, a threshold value can be set, for example, can be set threshold value be 3, judge the target whether there is in comprising
In class cluster of the target less than 3.
209: extracting the video frame where the target, provide security protection prompt.
If the target is present in number of targets less than in the class cluster of threshold value, indicating the behavioural characteristic of the target, there may be different
Often, therefore, it is desirable to determine the video frame where the target, the video frame number where the target can be specially extracted, by these
Video frame is marked and provides security protection prompt, to remind background work personnel to check video frame existing for the target,
Further determine that no when the target there are abnormal behaviours.
Above embodiment illustrates the specific implementation process of the security monitor for single target, and those skilled in the art can
To understand, other targets in video frame can also be handled using such as upper type, details are not described herein again.
In order to better understand to the present invention, Fig. 3 shows the specific embodiment of the present invention in practical applications, when this
When invention is applied to the scene being monitored to pedestrian, Video Monitoring Terminal can show monitored picture as shown in Figure 3, video
Include multiple target persons in picture, is illustrated below for the detailed process for prompting personage as the security protection of target.
Firstly, detecting to each human target in video, each human target is marked with boundary rectangle, it is each external
Rectangle just represents a human target, records the parameter information of the boundary rectangle of each human target.For example, for one of them
Human target, the parameter information in the i-th frame picture are (Xi, Yi, Wi, Hi), wherein X, Y indicate the external of the human target
The abscissa and ordinate (any vertex of boundary rectangle or central point can be taken) of rectangle in the video frame；W indicates boundary rectangle
Width, H indicate boundary rectangle height.
Next, intercepting the video frame of one section of video pictures according to certain frequency, such as can be according to the mark of 20 frame per second
Standard, the picture of interception one minute, then obtain the video pictures of 1200 frames.
Each human target for each of working as video frame for this section of video frame (occurs in addition to the human target
Other than first video frame), calculate the change of each frame of each target and the target boundary rectangle parameter information of former frame
Change value.For example, the parameter information of a human target in the i-th frame, and its temporal previous video frame, i.e., (i-1)-th
In identical human target parameter information (Xi-1, Yi-1, Wi-1, Hi-1) make the difference, the changing value of available each parameter information
(ΔXi, Δ Yi, Δ Wi, Δ Hi), this four-dimension array is defined as behavioural characteristic.Wherein, Δ Xi, Δ YiThe human target in X,
Displacement in Y-axis；ΔWiIndicate the width variation of human target, Δ HiIndicate the high variable quantity of human target.
For each human target, in the parameter information (X of the i-th framei, Yi, Wi, Hi) it can be understood as it one
Value in a four dimensions coordinate system in each dimension；In this way, frame each for this section of video is (in addition to the human target occurs
First video frame other than) each human target, one group of (Δ X can be obtainedi, Δ Yi, Δ Wi, Δ Hi) the four-dimension
Array, such as m human target are obtained with m group (Δ Xi, Δ Yi, Δ Wi, Δ Hi)；To whole frame in this section of video frame
The four-dimension array of whole human targets executes cluster, is divided into N number of cluster.
Because behavioural characteristic is convergent, therefore in general can for the personage of most normal pass
It is aggregated to the numerous class cluster of same or several numbers, and difference can be presented in a small number of abnormal behaviour persons, behavioural characteristic, it is thus possible to
Class cluster can be independently formed；In cluster result, the fewer one or several class clusters of number are chosen, such as a threshold can be set
Value extracts the wherein frame number where human target when the number in class cluster is less than threshold value, security protection prompt is provided, to remind
Backstage personnel check examine whether there is abnormal behaviour to the frame.
For the embodiment of the present invention, the interception frequency of video to be detected is less than video acquisition frequency, can be effectively reduced and is
The operation pressure of system reduces hardware cost；The present invention is only acquired the parameter information of boundary rectangle, parameter acquisition modes
It is simple and convenient, operand can be reduced；Meanwhile the present invention executes cluster as target person behavioural characteristic using many kinds of parameters
It calculates, the accuracy to abnormal behaviour judgement can be improved.
Further, as specific implementation of the invention, as shown in figure 4, the present invention also provides a kind of Behavior-based control features
The intelligent security guard prompt system of cluster specifically includes that object detection unit 401, target component determination unit 402, computing unit
403, cluster cell 404, prompt unit 405.
The object detection unit 401 can detect each target in each frame video pictures, and with external
Rectangle marked simultaneously tracks each target.
Each frame video pictures may have multiple targets, detect to multiple targets in video pictures, and will
The multiple targets detected are marked with boundary rectangle, each boundary rectangle just represents a target, in the present embodiment
Target can be personage, vehicle etc., and those skilled in the art can be configured according to specific requirements, here with no restrictions.
The target component determination unit 402 can recorde each target boundary rectangle in each frame video pictures
Parameter information.
It can be external as target using the transverse and longitudinal coordinate of the transverse and longitudinal coordinate on any vertex of target boundary rectangle or central point
The parameter information of rectangle, further, the parameter information of target boundary rectangle can also include that the width of target boundary rectangle is believed
Breath and elevation information.If there are multiple target boundary rectangles for present frame, to each target boundary rectangle recording parameters information.
The computing unit 403 can intercept one section of video frame, calculate separately the target boundary rectangle of each frame with it is previous
The changing value of each parameter information of the target boundary rectangle of the same target of frame forms one for each target in this section of video frame
A parameter variation value array is as behavioural characteristic.
It is specifically as follows, intercepts one section of video frame, such as can intercepts 20 seconds, 30 seconds, one minute etc.；According to step 102
The parameter information of the target boundary rectangle of each target of each frame of middle record, by each frame in this section of video (in addition to mesh
Mark other than first existing video frame) each target boundary rectangle parameter information same mesh corresponding with its former frame
The parameter information of target target boundary rectangle compares, the change of the parameter information of each target boundary rectangle of available each frame
Change value, specifically, the changing value can be by the parameter information of the target boundary rectangle of present frame and the same target of former frame
The parameter information of target boundary rectangle do difference operation.In this section of video frame, same target may persistently will appear in multiple
In frame, thus the target boundary rectangle parameter information of same target may to be multiple, the changing value of parameter may also be it is multiple,
Therefore a multi-Dimensional parameters changing value array, the behavior as the target can be formed for each of this section of video frame target
Characteristic information.
For example, parameter information includes the cross of central point for the target boundary rectangle of one of target in the i-th frame
Coordinate Xi, ordinate Yi, boundary rectangle width Wi, boundary rectangle height Hi, by the mesh of itself and the same target in the (i-1)-th frame
The corresponding parameter information for marking boundary rectangle compares, and obtains the changing value Δ X of parameter informationi, Δ Yi, Δ Wi, Δ Hi, can incite somebody to action
It forms a four-dimensional array (Δ Xi, Δ Yi, Δ Wi, Δ Hi), as the behavioural characteristic of the target, for this section of video
Each of target, one group of 4 D data can be obtained.
Further, video frame can be intercepted according to certain frequency in advance, interception frequency can be less than video
The frequency of acquisition, such as video pictures can be intercepted according to the frequency of 20 frame per second, since interception frequency is lower, obtain
Video frame quantity be not it is very huge, can reduce the treating capacity of data, improve data-handling efficiency.
The cluster cell 404 can execute the parameter variation value array of the target complete obtained from this section of video frame
Cluster calculation divides class cluster to the target in this section of video frame, chooses the class cluster that number of targets is less than threshold value.
For example, the multi-Dimensional parameters changing value array of the target complete in this section of video frame can be executed cluster calculation, draw
It is divided into N number of cluster, the similar target of behavioural characteristic can be generally gathered in the more class cluster of one or several targets, for behavior
Abnormal target, could be separately formed class cluster, therefore the class cluster less for destination number, the target for being included is very likely
There are abnormal behaviours.The less class cluster of destination number can be chosen to be analyzed, specifically, a threshold value can be set, work as class
When destination number is less than the threshold value in cluster, the video frame where the target in such cluster is extracted.
The prompt unit 405 provides security protection prompt for determining the place frame of the target in selected class cluster.
For independently forming the target of class cluster, it is most likely that there are abnormal behaviours, it is therefore desirable to extract these targets place
Video frame checked, be specifically as follows determine these targets where frame number, according to frame number extract video frame, provide security protection
Prompt, so as to backstage, personnel check in time, detect whether to exist abnormal.
In the present invention, be less than video acquisition frequency due to intercepting the frequency of video to be detected, obtained number of pictures compared with
It is few, the operation pressure of system can be effectively reduced；The present invention is only acquired the parameter information of boundary rectangle, and parameter obtains
Mode is simple and convenient, can reduce operand；Meanwhile the present invention using the coordinate information of target boundary rectangle as parameter in addition to believing
Breath, also using the width information of target boundary rectangle and elevation information as parameter information, therefore when carrying out cluster calculation, has
The accuracy rate of cluster calculation can be improved in the characteristic information of very abundant.
Based on method shown in above-mentioned Fig. 1, correspondingly, the embodiment of the invention also provides a kind of storage medium, the storage
An at least executable instruction is stored in medium, described execute instruction makes processor execute following steps: drawing to each frame video
Each target in face is detected, and marks each target with boundary rectangle；Record each mesh in each frame video pictures
Mark the parameter information of boundary rectangle；Intercept one section of video frame, calculate separately each frame target boundary rectangle and former frame it is same
The changing value of the parameter information of the target boundary rectangle of one target forms a parameter for each target in this section of video frame and becomes
Change value array；The parameter variation value array of the target complete obtained from this section of video frame is executed into cluster calculation, to this section of video
Target in frame divides class cluster, chooses the class cluster that number of targets is less than threshold value；Determine the place frame of the target in selected class cluster,
Provide security protection prompt.
Based on the embodiment of above-mentioned method as shown in Figure 1 and system as shown in Figure 4, the embodiment of the invention also provides one kind
The structural schematic diagram of memory and processor in computer equipment, as shown in figure 5, system of the invention may include one or more
A such as lower component: processor (processor) 510 and memory (memory) 520.Processor 510 may include one or
Multiple processing cores.Processor 510 is using the various pieces in various interfaces and the entire terminal of connection, by running or holding
Instruction, program, code set or the instruction set that row is stored in memory 520, and call the number being stored in memory 520
According to, execute terminal various functions and processing data.
Optionally, realize that above-mentioned each embodiment of the method provides when processor 510 executes the program instruction in memory 520
Behavior-based control feature clustering intelligent security guard reminding method.
Memory 520 may include random access memory ram, also may include read only memory ROM.Optionally, the storage
Device 520 includes non-transient computer-readable medium.Memory 520 can be used for store instruction, program, code, code set or refer to
Enable collection.Memory 520 may include storing program area and storage data area, wherein storing program area can store for realizing operation
The instruction of system, for the instruction of at least one function, for realizing instruction of above-mentioned each embodiment of the method etc.；Storing data
Area, which can be stored, uses created data etc. according to terminal.
Those of ordinary skill in the art will appreciate that realizing that all or part of the steps of above-described embodiment can pass through hardware
It completes, relevant hardware can also be instructed to complete by program, the program can store in a kind of computer-readable
In storage medium, storage medium mentioned above can be read-only memory, disk or CD etc..
The foregoing is merely the preferred embodiments of the application, not to limit the application, it is all in spirit herein and
Within principle, any modification, equivalent replacement, improvement and so on be should be included within the scope of protection of this application.
It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and this field skill
Art personnel can be designed alternative embodiment without departing from the scope of the appended claims.In the claims, it does not answer
Any reference symbol between parentheses is configured to limitations on claims.Word "comprising", which does not exclude the presence of, not to be listed in
Element or step in claim.Word "a" or "an" located in front of the element does not exclude the presence of multiple such members
Part.The present invention can be realized by means of including the hardware of several different elements and by means of properly programmed computer.
In the unit claims listing several devices, several in these devices, which can be through the same hardware branch, has
Body embodies.The use of word first, second, and third does not indicate any sequence.These words can be construed to title.
Claims (10)
1. a kind of intelligent security guard reminding method of the Behavior-based control feature clustering of Behavior-based control feature clustering, which is characterized in that institute
State method the following steps are included:
Step 1: detecting each target in each frame video pictures, and marks each target with boundary rectangle；
Step 2: the parameter information of each target boundary rectangle in each frame video pictures is recorded；
Step 3: one section of video frame of interception calculates separately the mesh of the target boundary rectangle of each frame and the same target of former frame
The changing value of the parameter information of boundary rectangle is marked, forms a parameter variation value array for each target in this section of video frame；
Step 4: the parameter variation value array of the target complete obtained from this section of video frame is executed into cluster calculation, which is regarded
Target in frequency frame divides class cluster, chooses the class cluster that number of targets is less than threshold value；
Step 5: determining the place frame of the target in selected class cluster, provides security protection prompt.
2. the method for intelligent security guard prompt as described in claim 1, which is characterized in that the target is the people in video pictures
Object.
3. the method for intelligent security guard prompt as described in claim 1, which is characterized in that in step 2, the target is external
The parameter information of rectangle is coordinate information of the target boundary rectangle in the video frame, and the coordinate information is expressed as outside target
The coordinate information of any vertex of rectangle or central point is connect, the coordinate information includes abscissa and ordinate.
4. the method for intelligent security guard prompt as claimed in claim 3, which is characterized in that the parameter of the target boundary rectangle is believed
Breath can also include the width information and elevation information of target boundary rectangle.
5. the method for intelligent security guard prompt as described in claim 1, which is characterized in that in step 3, one section of the interception
Video frame is specifically as follows: intercepting the video frame in one section of video according to certain frequency, the frequency is less than camera acquisition frequency
Rate.
6. the method for intelligent security guard prompt as described in claim 1, which is characterized in that in step 3, each frame
Changing value of the target boundary rectangle relative to the parameter information of the target boundary rectangle of the same target of former frame, specifically can be with
Are as follows: the parameter information of the target boundary rectangle of present frame parameter information corresponding with the same target boundary rectangle of former frame is done
Difference operation.
7. a kind of intelligent security guard of Behavior-based control feature clustering prompts equipment, which is characterized in that the system comprises:
Object detection unit: detecting each target in each frame video pictures, and marks each mesh with boundary rectangle
Mark；
Target component determination unit: the parameter information of each target boundary rectangle in each frame video pictures is recorded；
Computing unit: one section of video frame of interception calculates separately the target boundary rectangle of each frame and the same target of former frame
The changing value of the parameter information of target boundary rectangle forms a parameter variation value number for each target in this section of video frame
Group；
Cluster cell: the parameter variation value array of the target complete obtained from this section of video frame is executed into cluster calculation, to the section
Target in video frame divides class cluster, chooses the class cluster that number of targets is less than threshold value；
Prompt unit: determining the place frame of the target in selected class cluster, provides security protection prompt.
8. the system of intelligent security guard prompt as claimed in claim 7, which is characterized in that the target in the object detection unit
For the personage in video pictures.
9. the method for intelligent security guard prompt as claimed in claim 7, which is characterized in that in target component determination unit, institute
The parameter information for stating target boundary rectangle is coordinate information of the target boundary rectangle in the video frame, the coordinate information table
It is shown as the coordinate information of any vertex of target boundary rectangle or central point, the coordinate information includes abscissa and ordinate.
10. the method for intelligent security guard prompt as claimed in claim 9, which is characterized in that in target component determination unit, institute
The parameter information for stating target boundary rectangle can also include the width information and elevation information of target boundary rectangle.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910478050.3A CN110378200A (en) | 2019-06-03 | 2019-06-03 | A kind of intelligent security guard prompt apparatus and method for of Behavior-based control feature clustering |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910478050.3A CN110378200A (en) | 2019-06-03 | 2019-06-03 | A kind of intelligent security guard prompt apparatus and method for of Behavior-based control feature clustering |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110378200A true CN110378200A (en) | 2019-10-25 |
Family
ID=68249716
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910478050.3A Pending CN110378200A (en) | 2019-06-03 | 2019-06-03 | A kind of intelligent security guard prompt apparatus and method for of Behavior-based control feature clustering |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110378200A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110991461A (en) * | 2019-10-29 | 2020-04-10 | 重庆特斯联智慧科技股份有限公司 | Intelligent extraction key security target excavation method and system |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101389004A (en) * | 2007-09-13 | 2009-03-18 | 中国科学院自动化研究所 | Moving target classification method based on on-line study |
| CN102156880A (en) * | 2011-04-11 | 2011-08-17 | 上海交通大学 | Method for detecting abnormal crowd behavior based on improved social force model |
| CN103942533A (en) * | 2014-03-24 | 2014-07-23 | 河海大学常州校区 | Urban traffic illegal behavior detection method based on video monitoring system |
| CN103996051A (en) * | 2014-05-12 | 2014-08-20 | 上海大学 | Method for automatically detecting abnormal behaviors of video moving object based on change of movement features |
| CN105741326A (en) * | 2016-03-21 | 2016-07-06 | 西安电子科技大学 | Target tracking method for video sequence based on clustering fusion |
| CN106127814A (en) * | 2016-07-18 | 2016-11-16 | 四川君逸数码科技股份有限公司 | A kind of wisdom gold eyeball identification gathering of people is fought alarm method and device |
| CN106203274A (en) * | 2016-06-29 | 2016-12-07 | 长沙慧联智能科技有限公司 | Pedestrian's real-time detecting system and method in a kind of video monitoring |
| CN106469276A (en) * | 2015-08-19 | 2017-03-01 | 阿里巴巴集团控股有限公司 | The kind identification method of data sample and device |
| CN108009473A (en) * | 2017-10-31 | 2018-05-08 | 深圳大学 | Based on goal behavior attribute video structural processing method, system and storage device |
| CN108322363A (en) * | 2018-02-12 | 2018-07-24 | 腾讯科技（深圳）有限公司 | Propelling data abnormality monitoring method, device, computer equipment and storage medium |
| CN109145934A (en) * | 2017-12-22 | 2019-01-04 | 北京数安鑫云信息技术有限公司 | User behavior data processing method, medium, equipment and device based on log |
| CN109829382A (en) * | 2018-12-30 | 2019-05-31 | 北京宇琪云联科技发展有限公司 | The abnormal object early warning tracing system and method for Behavior-based control feature intelligent analysis |
-
2019
- 2019-06-03 CN CN201910478050.3A patent/CN110378200A/en active Pending
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101389004A (en) * | 2007-09-13 | 2009-03-18 | 中国科学院自动化研究所 | Moving target classification method based on on-line study |
| CN102156880A (en) * | 2011-04-11 | 2011-08-17 | 上海交通大学 | Method for detecting abnormal crowd behavior based on improved social force model |
| CN103942533A (en) * | 2014-03-24 | 2014-07-23 | 河海大学常州校区 | Urban traffic illegal behavior detection method based on video monitoring system |
| CN103996051A (en) * | 2014-05-12 | 2014-08-20 | 上海大学 | Method for automatically detecting abnormal behaviors of video moving object based on change of movement features |
| CN106469276A (en) * | 2015-08-19 | 2017-03-01 | 阿里巴巴集团控股有限公司 | The kind identification method of data sample and device |
| CN105741326A (en) * | 2016-03-21 | 2016-07-06 | 西安电子科技大学 | Target tracking method for video sequence based on clustering fusion |
| CN106203274A (en) * | 2016-06-29 | 2016-12-07 | 长沙慧联智能科技有限公司 | Pedestrian's real-time detecting system and method in a kind of video monitoring |
| CN106127814A (en) * | 2016-07-18 | 2016-11-16 | 四川君逸数码科技股份有限公司 | A kind of wisdom gold eyeball identification gathering of people is fought alarm method and device |
| CN108009473A (en) * | 2017-10-31 | 2018-05-08 | 深圳大学 | Based on goal behavior attribute video structural processing method, system and storage device |
| CN109145934A (en) * | 2017-12-22 | 2019-01-04 | 北京数安鑫云信息技术有限公司 | User behavior data processing method, medium, equipment and device based on log |
| CN108322363A (en) * | 2018-02-12 | 2018-07-24 | 腾讯科技（深圳）有限公司 | Propelling data abnormality monitoring method, device, computer equipment and storage medium |
| CN109829382A (en) * | 2018-12-30 | 2019-05-31 | 北京宇琪云联科技发展有限公司 | The abnormal object early warning tracing system and method for Behavior-based control feature intelligent analysis |
Non-Patent Citations (1)
| Title |
|---|
| 李祥民等: "基于聚类的目标行为异常检测", 《计算机与网络》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110991461A (en) * | 2019-10-29 | 2020-04-10 | 重庆特斯联智慧科技股份有限公司 | Intelligent extraction key security target excavation method and system |
| CN110991461B (en) * | 2019-10-29 | 2021-05-25 | 重庆特斯联智慧科技股份有限公司 | Intelligent extraction key security target excavation method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103761748B (en) | Anomaly detection method and device | |
| CN103093212B (en) | The method and apparatus of facial image is intercepted based on Face detection and tracking | |
| CN106650584B (en) | Flame detecting method and system | |
| CN102214359B (en) | Target tracking device and method based on hierarchic type feature matching | |
| CN105787472B (en) | A kind of anomaly detection method based on the study of space-time laplacian eigenmaps | |
| CN108009473A (en) | Based on goal behavior attribute video structural processing method, system and storage device | |
| CN105376255A (en) | Android platform intrusion detection method based on K-means cluster | |
| JP2012518846A (en) | System and method for predicting abnormal behavior | |
| CN107506734A (en) | One kind of groups unexpected abnormality event detection and localization method | |
| CN108021846A (en) | A kind of face identification method and device | |
| CN104820824A (en) | Local abnormal behavior detection method based on optical flow and space-time gradient | |
| CN110263633A (en) | The personnel that are involved in drug traffic based on space time correlation detect method for early warning, system and storage medium | |
| CN112163572A (en) | Method and device for identifying object | |
| CN106228709A (en) | It is a kind of that wisdom gold eyeball identification is single adds paper money alarm method and device | |
| CN106792883A (en) | Sensor network abnormal deviation data examination method and system | |
| CN103761516A (en) | ATM abnormal human face detection method based on video monitoring | |
| CN115861915A (en) | Fire fighting access monitoring method, fire fighting access monitoring device and storage medium | |
| CN110378200A (en) | A kind of intelligent security guard prompt apparatus and method for of Behavior-based control feature clustering | |
| CN109146913B (en) | Face tracking method and device | |
| CN112580531B (en) | Identification detection method and system for true and false license plates | |
| CN109784253A (en) | A kind of detection method of bicycle user abnormal behaviour | |
| CN116682162A (en) | Robot detection algorithm based on real-time video stream | |
| Liu et al. | Analysis and design of public places crowd stampede early-warning simulating system | |
| Wateosot et al. | Group activity recognition with an interaction force based on low‐level features | |
| CN111369394B (en) | Scenic spot passenger flow volume statistical evaluation system and method based on big data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191025 |
|
| RJ01 | Rejection of invention patent application after publication |