CN109309811A - A kind of throwing object in high sky detection system based on computer vision and method - Google Patents
A kind of throwing object in high sky detection system based on computer vision and method Download PDFInfo
- Publication number
- CN109309811A CN109309811A CN201811015244.1A CN201811015244A CN109309811A CN 109309811 A CN109309811 A CN 109309811A CN 201811015244 A CN201811015244 A CN 201811015244A CN 109309811 A CN109309811 A CN 109309811A
- Authority
- CN
- China
- Prior art keywords
- high sky
- throwing object
- image data
- target
- moving target
- 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
- H04N7/181—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast for receiving images from a plurality of remote sources
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N5/00—Details of television systems
- H04N5/14—Picture signal circuitry for video frequency region
- H04N5/144—Movement detection
Abstract
The invention discloses a kind of throwing object in high sky detection system based on computer vision and methods, comprising: video image acquisition unit, for being monitored in real time to cell and obtaining image data comprehensively；Network transmitting unit is used for transmission image data；Moving object detection unit is analyzed and processed the image data of acquisition for receiving the image data from network transmitting unit, determines the moving target in image data；Center processing unit obtains the feature and motion profile of moving target, and judge whether it belongs to throwing object in high sky, obtains the position of dishing out for belonging to throwing object in high sky target, be related to community security system field for being tracked to moving target.The present invention is applied widely, and versatile, detection effect is good.
Description
Technical field
The present invention relates to community security system fields, are specifically related to a kind of throwing object in high sky detection based on computer vision
System and method for.
Background technique
Throwing object in high sky is referred to as " pain for being suspended in city overhead ", and throwing object in high sky behavior all the time is concerned, as city
While city's uncivil behavior, social danger brought by it is also very big, high from 12 meters or so by scientific algorithm and test
Dish out 30 grams of egg of floor is pounded and can make swollen packet with it, thrown from 54 meters or so it is lower the skull of people can be broken, from 75
Rice or so can make one to die instantly under throwing, and cause the injury that can not be retrieved.Since the implementation place of this uncivil behavior is more
Position high-altitude floor, rare witness, the parabolic time is short or even someone deliberately conceals figure in parabolic, is difficult law enforcement agency
The legal liabilities of shooter are investigated, such event is commonplace, remains incessant after repeated prohibition.Throwing object in high sky behavior destroys living environment, danger
The life security for having done harm to the people, has invaded the legitimate rights and interests of the people, and management regulation is related to the stabilization of social security.
Currently, supervision throwing object in high sky behavior is mainly or by the not timing patrol of resident's spontaneous organization and in throwing object in high sky
Put up the methods of caution slogan in recurrent location；But the method for not timing patrol can not monitor throwing object in high sky completely
Behavior, and expend a large amount of manpowers；The method for putting up caution slogan is easily ignored, can not cause enough attention, the effect played
Fruit is extremely limited.Under the premise of herein, in order to further strengthen the prevention of very harmful uncivil behavior this to throwing object in high sky,
We need to build a kind of throwing object in high sky detection system based on computer vision, throwing object in high sky event generation after, convenient for into
Row post-mordem forensics and the legal liabilities for investigating related personnel, the mode based on civilization publicity, supplemented by monitoring management, effectively
Prevent the generation of throwing object in high sky behavior.
Summary of the invention
The purpose of the invention is to overcome the shortcomings of above-mentioned background technique, a kind of base for being convenient for post-mordem forensics is provided
In the throwing object in high sky detection system and method for computer vision.
The present invention provides a kind of throwing object in high sky detection system based on computer vision, comprising:
Video image acquisition unit, for being monitored in real time to cell and obtaining image data comprehensively；
Network transmitting unit is used for transmission image data；
Moving object detection unit, for receiving the image data from network transmitting unit, to the image data of acquisition
It is analyzed and processed, determines the moving target in image data；
Center processing unit obtains the feature and motion profile of moving target, and sentence for being tracked to moving target
Breaking, whether it belongs to throwing object in high sky, obtains the position of dishing out for belonging to throwing object in high sky target.
Based on the above technical solution, the video image acquisition unit includes multiple web cameras and a number
According to stream processing unit, multiple web cameras are mounted in cell at different location, and the Data Stream Processing unit is used for
It obtains the video data stream of each web camera and image data is decoded as using H265 decoding technique.
Based on the above technical solution, the throwing object in high sky detection system based on computer vision further includes video
Storage unit, the video storage unit are used to cache the movement for belonging to throwing object in high sky target by central processing unit judgement
Track video flow data.
Based on the above technical solution, the video storage unit leads to the image data for belonging to throwing object in high sky target
Cross H265 compression method boil down to video stream data.
Based on the above technical solution, the video image acquisition unit, network transmitting unit, moving object detection
Data exchange is carried out using buffer area between unit, center processing unit.
The present invention also provides a kind of throwing object in high sky detection methods based on computer vision, comprising the following steps:
1) Image Acquisition；
Real time monitoring cell obtains parabolic video stream data comprehensively, and video stream data is decoded as image data；
2) moving object detection；
Analysis processing image data, determines the moving target in image data；
3) tracking moving object；
Each moving target is tracked, the feature and motion profile of moving target are obtained；
4) throwing object in high sky determines；
Analysis determines whether moving target belongs to throwing object in high sky, obtains the position of dishing out for belonging to throwing object in high sky target.
Based on the above technical solution, in step 2), moving object detection step are as follows:
(1) in the first frame of image data, a sample set is stored for each pixel, adjacent point around pixel
Pixel value is exactly the sampled value of sample set；
(2) sampled value of the pixel value of the pixel of later each frame and sample set is compared, if the difference is less than threshold
Value, then it is assumed that two o'clock is close；
(3) all sample sets are successively traversed, number similar in pixel and sample set can be obtained, if number is greater than threshold value,
Then think that the pixel is background dot, otherwise it is assumed that the pixel is foreground point；
(4) obtained foreground point is screened, undesirable object is screened out according to the perimeter L of object and area S
Body carries out burn into expansion, connectivity analysis removal detection noise to testing result, finally obtains the movement of minimum 7*7 pixel
Target, and extract the feature of moving target.
Based on the above technical solution, in step 3), tracking moving object step are as follows:
(1) obtaining step 2) in the central point of moving target and the stingy figure of moving target；
(2) environment of first frame initialization Kalman filtering algorithm, and all moving targets are defaulted as what needs were tracked
Target；
(3) each moving target of subsequent each frame, is matched by Hungary Algorithm with existing track, if matching
Success, then be added into corresponding track；If it fails to match, predict this track current by Kalman filtering algorithm
The present position of frame；
(4) if certain track continuous several times are matched to moving object not successfully, that is, it is pre- to continue through Kalman filtering algorithm
The number of the tracing point measured is greater than particular value A, then it is assumed that this track terminates；
(5) if the number of the point of certain track is greater than threshold value B, then it is assumed that the track is effective, otherwise it is assumed that it is invalid.
Based on the above technical solution, in step 4), the decision condition of throwing object in high sky is as follows:
(1) if what the value of the ordinate of movement objective orbit persistently increased；
(2) the number M of throwing object in high sky movement objective orbit point is greater than preset threshold value T, i.e. M > T；
(3) throwing object in high sky movement objective orbit terminates from the boundary of monitoring range；
(4) projectile motion is accelerated in beginning；
(5) in throwing object in high sky moving target dropping process, the difference numerical of ordinate will not be too small always.
Based on the above technical solution, in step 4), obtain belong to throwing object in high sky target dish out position the step of
Are as follows:
(1) it is directed to every track, all the points on track are fitted using least square method, is done using frame number from change
It measures, the numerical value on the direction y obtains a function as dependent variable, fitting；
(2) according to the function, from first point of existing track toward rewinding, until the numerical value in the direction y of continuous two points
Difference is less than threshold value A, then it is assumed that reaches throw point, current point is saved as and is dished out to a track.
Compared with prior art, advantages of the present invention is as follows:
(1) throwing object in high sky detection system based on computer vision of the invention makes full use of computer vision technique, needle
The actual conditions that throwing object in high sky occurs detect, and reinforce the safeguard management of cell, improve cell environment, for Small object
Moving object has preferable detection effect, can adapt to target and is blocked by of short duration or situations such as missing inspection, can adapt to background ring
Slowly varying situation occurs for border, can remove the non-projectile motion target in part in tracing process, applied widely, versatile,
Detection effect is good.
(2) throwing object in high sky detection system based on computer vision of the invention carries out modular division to system, will be each
Buffer chain is utilized between a module, is sufficiently reduced the coupling between modules, and can reduce the loss of data, is mentioned
The stability and robustness of high system.
(3) throwing object in high sky detection method based on computer vision of the invention passes through the video of the monitoring scene obtained,
Real-time detection Small object Moving Objects, are then tracked moving target, and can remove a part of non-projectile motion target,
It is carried out curve fitting by the motion profile to moving target, in conjunction with throwing object in high sky sports rule, judges the movement that detected
Whether target is throwing object in high sky, and the method makes full use of the sports rule of throwing object in high sky, and algorithm principle is easily understood, and realizes process
Simply, determine that effect is more accurate.
Detailed description of the invention
Fig. 1 is a kind of designed holder composition of the throwing object in high sky detection system of computer vision of the embodiment of the present invention；
Fig. 2 is the Producer-consumer problem synchro-control model data buffering schematic diagram of the embodiment of the present invention；
Fig. 3 is a kind of operation data stream signal of the throwing object in high sky detection method of computer vision of the embodiment of the present invention
Figure；
Fig. 4 is the refinement process of the ViBe background modeling moving target detecting method based on OpenCL of the embodiment of the present invention
Schematic diagram；
Fig. 5 is the two-dimentional ViBe Model Matching schematic diagram of the embodiment of the present invention；
Fig. 6 is the Kalman filtering tracking moving object side based on throwing object in high sky movement priori knowledge of the embodiment of the present invention
The refinement flow diagram of method.
Specific embodiment
With reference to the accompanying drawing and specific embodiment the present invention is described in further detail.
Shown in Figure 1, the embodiment of the present invention provides a kind of throwing object in high sky detection system based on computer vision, packet
It includes:
Video image acquisition unit, for being monitored in real time to cell and obtaining image data comprehensively；
Network transmitting unit is used for transmission image data；
Moving object detection unit, for receiving the image data from network transmitting unit, to the image data of acquisition
It is analyzed and processed, determines the moving target in image data；
Center processing unit obtains the feature and motion profile of moving target, and sentence for being tracked to moving target
Breaking, whether it belongs to throwing object in high sky, obtains the position of dishing out for belonging to throwing object in high sky target.
Throwing object in high sky detection system based on computer vision of the invention makes full use of computer vision technique, for height
The actual conditions that empty parabolic occurs are detected, and the safeguard management of cell is reinforced, and improve cell environment, the movement for Small object
Object has preferable detection effect, can adapt to target and is blocked by of short duration or situations such as missing inspection, can adapt to background environment hair
Slowly varying situation is given birth to, can remove the non-projectile motion target in part in tracing process, applied widely, versatile, detection
Effect is good.
On the basis of the technical solution of above-described embodiment, further, the video image acquisition unit includes multiple nets
Network video camera and a data stream processing unit, multiple web cameras are mounted in cell at different location, the number
It is used to obtain the video data stream of each web camera according to stream processing unit and image data is decoded as using H265 decoding technique；
The arrangement of monitor camera is mounted on the lower section in monitoring building with installation using web camera, and camera is mounted on apart from building 30
Rice place and mounting height are 3 meters, with this condition using suitable number of cameras, each camera according to suitable setting angle into
Row installation, it is ensured that the ratio of the size 1:1 of article is imaged in object actual size and monitoring picture, if both objects is true big
Small is N1*N2 (cm), then the size of the object shot in video is N1*N2 (pixel), and is connected and passed by network transmitting unit
Defeated monitoring data；Video data stream obtains, and obtains the video compressed by the IP address of camera, port numbers, user name, password
Data flow；Data stream, using H265 decoding technique by the video data stream decoding of acquisition be a frame frame picture.
On the basis of the technical solution of above-described embodiment, further, the throwing object in high sky inspection based on computer vision
Examining system further includes video storage unit, and the video storage unit belongs to height by central processing unit judgement for caching
The motion profile video stream data of empty parabolic target；Scene when reproducible throwing object in high sky, convenient for subsequent verification.
On the basis of the technical solution of above-described embodiment, further, the video storage unit will belong to throwing object in high sky
The image data of target passes through H265 compression method boil down to video stream data；When determining throwing object in high sky occur, system is delayed
The monitoring picture for a period of time deposited realizes throwing object in high sky key view by the compression that H265 compression method realizes picture to video
The preservation of frequency, and the time that can be occurred according to parabolic, the information searchings such as position are played back, to determine going out for throwing object in high sky
Hair point.
On the basis of the technical solution of above-described embodiment, further, the video image acquisition unit, network transmission list
Member utilizes buffer area progress data exchange between moving object detection unit, center processing unit；Using each module of system as
Individual sub thread, centre are attached by buffering queue, and buffering queue does thread synchronization control, and control mode is using production
Person-Consumer model.Such as the synchronously control between video image acquisition unit and moving object detection unit, by giving system
Image pick-up signal is sent, Image Acquisition thread is waken up, obtains monitoring video flow and to decoding video stream, acquisition video original graph
Piece is linked into buffer area tail of the queue, and changes critical zone coherent signal amount；When there is image data in buffer area, moving target is just waken up
Thread is detected, picture is taken out from team's head and is further processed.Make there can be the short time between each module of system using this mode
Tolerance.But this time be also it is conditional, the setting of the size of buffer area determines the size of tolerance time.If system some
Thread process speed is excessively slow, causes its relevant buffer data cumulative, when close to buffer area total size, is
System hangs up the thread of its previous module, and feedback assembly line reduces speed command, prevents buffer overflow system crash.Pass through
Such scheme sufficiently reduces the coupling between modules, and can reduce the loss of data, improve system stability and
Robustness.
Referring to fig. 2 to shown in Fig. 6, the embodiment of the present invention also provides a kind of throwing object in high sky detection side based on computer vision
Method, comprising the following steps:
1) Image Acquisition；
Real time monitoring cell obtains parabolic video stream data comprehensively, and video stream data is decoded as image data；
2) moving object detection；
Analysis processing image data, determines the moving target in image data；
3) tracking moving object；
Each moving target is tracked, the feature and motion profile of moving target are obtained；
4) throwing object in high sky determines；
Analysis determines whether moving target belongs to throwing object in high sky, obtains the position of dishing out for belonging to throwing object in high sky target.
Throwing object in high sky detection method based on computer vision of the invention is by the video of the monitoring scene obtained, in real time
Small object Moving Objects are detected, then moving target is tracked, and a part of non-projectile motion target can be removed, is passed through
It carries out curve fitting to the motion profile of moving target, in conjunction with throwing object in high sky sports rule, judges the moving target that detected
It whether is throwing object in high sky, the method makes full use of the sports rule of throwing object in high sky, and algorithm principle is easily understood, and realizes process letter
It is single, determine that effect is more accurate.
On the basis of the technical solution of above-described embodiment, further, in step 2), moving object detection step are as follows:
(1) in the first frame of image data, a sample set is stored for each pixel, adjacent point around pixel
Pixel value is exactly the sampled value of sample set；
(2) sampled value of the pixel value of the pixel of later each frame and sample set is compared, if the difference is less than threshold
Value, then it is assumed that two o'clock is close；
(3) all sample sets are successively traversed, number similar in pixel and sample set can be obtained, if number is greater than threshold value,
Then think that the pixel is background dot, otherwise it is assumed that the pixel is foreground point；
(4) obtained foreground point is screened, undesirable object is screened out according to the perimeter L of object and area S
Body carries out burn into expansion, connectivity analysis removal detection noise to testing result, finally obtains the movement of minimum 7*7 pixel
Target, and extract the feature of moving target.
After the moving object detection is by carrying out gray processing and median filtering pretreatment to original image, using being based on
The moving target detecting method of the ViBe background modeling of OpenCL carries out burn into expansion to testing result, connectivity analysis removes
Detection noise, can finally obtain the moving target of minimum 7*7 pixel, and extract the feature of moving target.The signal of its process
Figure is as shown in Figure 3.
The realization of ViBe algorithm is broadly divided into three modules: background model initializing, Model Matching, background model update.
Firstly, only needing to utilize input video sequence first frame picture or a few frame pictures when background model initializing.It is right
Each pixel in picture randomly chooses the pixel value of this point or the pixel value of neighborhood as model.Such as in image
(x, y) pixel X (x, y), if following formula method is from its 8 neighborhood N8(X) N number of sample is randomly selected in join as model initialization
Number.
M (X)={ vi(y)|y∈N8(X) }, i=1,2 ..., N
Then, Model Matching is carried out to the image of input, it is similar to background model by calculating input image pixel value
Degree detection moving target.Assuming that present incoming frame is Xt, the pixel value at pixel (x, y) is Xt(x,y).As shown in figure 4, fixed
Justice one is with XtCentered on (x, y), R be radius sphere SR(Xt(x, y)), SR(Xt(x, y)) indicate all and Xt(x, y) distance
The set of point less than R falls in S with M (X)R(Xt(x, y)) the number m of sample indicates X in regiont(x, y) and background model M
(X) similarity, for given threshold value TmIf: m < Tm, then determine Xt(x, y) is prospect；If m >=TmThen determine Xt
(x, y) is background.It is shown below, dist (Xt(x,y),Mi(X)) X is indicatedt(x, y) and Mi(X) distance.
(Tm: threshold value R: radius)
Finally, enabling model preferably to describe current time to adapt to the variation of the background of model in monitoring scene
Background needs to be updated model.ViBe algorithm is according to current frame pixel value XtCarry out the background model M of renewable time t-1
(Xt-1), the probability that each background dot 1/ φ of probability having the same has come more new model, while utilizing pixel value spatial
Characteristic also has identical 1/ φ of probability to be updated model neighborhood point.Certain in continuous a period of time, in t moment model
The probability that a sample does not update is (N-1)/N, then calculation method of the sample again without changed probability after the dt time
Such as following formula:
ViBe moving object detection is largely the calculating operation of the pixel of image, in order to make system have response faster
Speed can do the improvement of parallelization to image procossing ViBe algorithm here.Using OpenCL programmed method to Algorithm parallelization
Processing, make full use of the computation capability that GPU is powerful and cooperate with CPU, more efficient using hardware-efficient
Large-scale, high degree of parallelism calculating in ViBe algorithm is completed, realizes system acceleration function.OpenCL programming is utilized GPU's
Feature carries out multi-threading parallel process and realizes ViBe algorithm, and the algorithm after parallelization has significantly in the high-resolution image of processing
Advantage.
On the basis of the technical solution of above-described embodiment, further, in step 3), tracking moving object step are as follows:
(1) obtaining step 2) in the central point of moving target and the stingy figure of moving target；
(2) environment of first frame initialization Kalman filtering algorithm, and all moving targets are defaulted as what needs were tracked
Target；
(3) each moving target of subsequent each frame, is matched by Hungary Algorithm with existing track, if matching
Success, then be added into corresponding track；If it fails to match, predict this track current by Kalman filtering algorithm
The present position of frame；
(4) if certain track continuous several times are matched to moving object not successfully, that is, it is pre- to continue through Kalman filtering algorithm
The number of the tracing point measured is greater than particular value A, then it is assumed that this track terminates；
(5) if the number of the point of certain track is greater than threshold value B, then it is assumed that the track is effective, otherwise it is assumed that it is invalid.
The tracking moving object is transported by the feature of the moving target of the detection of extraction using based on throwing object in high sky
The Kalman filtering method for tracking moving object of dynamic priori knowledge, the position of predicted motion target in the next frame determines target
The region matched finally realizes the matching of moving target, flow diagram such as Fig. 5 by Hungary's matching process.
Kalman filter realizes that target tracking is divided into three steps: initialized card Thalmann filter first substantially, by mesh
Target position and speed is assigned to Kalman filter, and records the time of present frame；Then state estimation is carried out, using current
The time interval of frame and previous frame carries out Kalman Prediction, predicts position and the motion state of the target of present frame, and the field of search
Best match in domain, obtains position and the state of motion value of real goal, while recording the current time, innovative herein
The priori knowledge for increasing throwing object in high sky movement, if there is track continuously non-fall time tnumT is greater than threshold value T, i.e. tnum> T, then table
Show that this moving target is non-projectile motion, directly remove this moving target, as following formula indicates；State update is finally carried out, is used
Time of day vector updates the state of Kalman filter as the input of Kalman filter.Kalman filter formulation mainly has
It is five, as follows:
Kalman Prediction equation
State equation:
Covariance predictive equation:
Kalman's update equation
The filtering gain of Kalman:
Modify state vector:
Modify covariance matrix:
Hungary Algorithm searches its augmenting path by the bipartite graph constituted to each point to be matched and obtains maximum matched mistake
Journey is broadly divided into three steps.
It (1) is sky by the matching figure M of bipartite graph G；
(2) mono- augmenting path P of bipartite graph G is found out, and is obtained more by augmenting path P and former bipartite graph G xor operation
The matching figure M' of big bipartite graph replaces M；
(3) step (2) are repeated, continues to search augmenting path.It is searched when all augmenting paths and completes then to stop step (2),
Matching is completed.
On the basis of the technical solution of above-described embodiment, further, in step 4), the decision condition of throwing object in high sky is such as
Under:
(1) if what the value of the ordinate of movement objective orbit persistently increased；
(2) the number M of throwing object in high sky movement objective orbit point is greater than preset threshold value T, i.e. M > T；
(3) throwing object in high sky movement objective orbit terminates from the boundary of monitoring range；
(4) projectile motion is accelerated in beginning；
(5) in throwing object in high sky moving target dropping process, the difference numerical of ordinate will not be too small always.
The throwing object in high sky determines, by throwing object in high sky motion profile feature, proposes the throwing object in high sky based on curve matching
Behavior determination method, throwing object in high sky be downwards, accelerate, parabolic exist in video a period of time, parabolic from monitoring model
Surrounding edge circle termination determines whether moving target is throwing object in high sky, and throwing object in high sky realizes that parabolic is dished out position by the fitting to curve
Set determination.The decision condition of throwing object in high sky determines the accuracy of judgement, according to the movement characteristic of throwing object in high sky, is mainly utilized
The essential condition that following parabolic feature judges as parabolic:
(1) motion profile of throwing object in high sky moves downward always, therefore the value of the ordinate of movement objective orbit is to hold
Continuous increase；
(2) the number M of throwing object in high sky movement objective orbit point will meet the threshold value T of setting, i.e. M > T, and moving target is being supervised
Movement in range is controlled, movement can move the regular hour surely, therefore the number of the tracing point detected has certain limitation；
(3) throwing object in high sky movement objective orbit terminates from the boundary of monitoring range, and projectile motion will not somewhere stop
Stay, from some borderland can go out monitoring range, while borderline region will not be coboundary region, only can from it is left and right and under
Borderline region；
(4) projectile motion is substantially being accelerated in beginning, but is influenced by environment etc., not can avoid meeting
There is detection target not slowing down and partial dot at the uniform velocity, but accelerates on the whole, therefore Set scale threshold value is to remove
Always at the uniform velocity or slow down moving target；
(5) in throwing object in high sky moving target dropping process, the difference numerical of ordinate will not be too small always, therefore can go
The noise being almost overlapped except ordinate.
On the basis of the technical solution of above-described embodiment, further, in step 4), acquisition belongs to throwing object in high sky target
Dish out position the step of are as follows:
(1) it is directed to every track, all the points on track are fitted using least square method, is done using frame number from change
It measures, the numerical value on the direction y obtains a function as dependent variable, fitting；
(2) according to the function, from first point of existing track toward rewinding, until the numerical value in the direction y of continuous two points
Difference is less than threshold value A, then it is assumed that reaches throw point, current point is saved as and is dished out to a track.
The throw point of traceable throwing object in high sky accordingly finds out accident responsibility people, is convenient for post-mordem forensics and investigates correlation
The legal liabilities of personnel are conducive to the generation again for preventing throwing object in high sky event, have ensured the life security of the people, maintained people
The legitimate rights and interests of the people are conducive to the stability for improving social security.
Those skilled in the art can carry out various modifications to the embodiment of the present invention and modification, if these modifications and change
For type within the scope of the claims in the present invention and its equivalent technologies, then these modifications and variations are also in protection scope of the present invention
Within.
The prior art that the content being not described in detail in specification is known to the skilled person.
Claims (10)
1. a kind of throwing object in high sky detection system based on computer vision characterized by comprising
Video image acquisition unit, for being monitored in real time to cell and obtaining image data comprehensively；
Network transmitting unit is used for transmission image data；
Moving object detection unit carries out the image data of acquisition for receiving the image data from network transmitting unit
Analysis processing, determines the moving target in image data；
Center processing unit obtains the feature and motion profile of moving target, and judge it for being tracked to moving target
Whether belong to throwing object in high sky, obtains the position of dishing out for belonging to throwing object in high sky target.
2. a kind of throwing object in high sky detection system based on computer vision as described in claim 1, it is characterised in that: the view
Frequency image acquisition units include multiple web cameras and a data stream processing unit, and multiple web cameras are mounted on
In cell at different location, the Data Stream Processing unit is used to obtain the video data stream of each web camera and using H265
Decoding technique is decoded as image data.
3. a kind of throwing object in high sky detection system based on computer vision as claimed in claim 2, it is characterised in that: the base
It further include video storage unit in the throwing object in high sky detection system of computer vision, the video storage unit is for caching by institute
State the motion profile video stream data that central processing unit judgement belongs to throwing object in high sky target.
4. a kind of throwing object in high sky detection system based on computer vision as claimed in claim 3, it is characterised in that: the view
The image data for belonging to throwing object in high sky target is passed through H265 compression method boil down to video stream data by frequency storage unit.
5. a kind of throwing object in high sky detection system based on computer vision as described in claim 1, it is characterised in that: the view
Frequency image acquisition units, moving object detection unit, are counted between center processing unit network transmitting unit using buffer area
According to exchange.
6. a kind of throwing object in high sky detection method based on computer vision, which comprises the following steps:
1) Image Acquisition；
Real time monitoring cell obtains parabolic video stream data comprehensively, and video stream data is decoded as image data；
2) moving object detection；
Analysis processing image data, determines the moving target in image data；
3) tracking moving object；
Each moving target is tracked, the feature and motion profile of moving target are obtained；
4) throwing object in high sky determines；
Analysis determines whether moving target belongs to throwing object in high sky, obtains the position of dishing out for belonging to throwing object in high sky target.
7. a kind of throwing object in high sky detection method based on computer vision as claimed in claim 6, it is characterised in that: step 2)
In, moving object detection step are as follows:
(1) in the first frame of image data, a sample set is stored for each pixel, the pixel of adjacent point around pixel
Value is exactly the sampled value of sample set；
(2) sampled value of the pixel value of the pixel of later each frame and sample set is compared, if the difference is less than threshold value,
Think that two o'clock is close；
(3) all sample sets are successively traversed, number similar in pixel and sample set can be obtained, if number is greater than threshold value, are recognized
It is background dot for the pixel, otherwise it is assumed that the pixel is foreground point；
(4) obtained foreground point is screened, undesirable object is screened out according to the perimeter L of object and area S, it is right
Testing result carries out burn into expansion, connectivity analysis removes detection noise, finally obtains the moving target of minimum 7*7 pixel,
And extract the feature of moving target.
8. a kind of throwing object in high sky detection method based on computer vision as claimed in claim 7, it is characterised in that: step 3)
In, tracking moving object step are as follows:
(1) obtaining step 2) in the central point of moving target and the stingy figure of moving target；
(2) environment of first frame initialization Kalman filtering algorithm, and all moving targets are defaulted as the target that needs are tracked；
(3) then each frame each moving target, matched by Hungary Algorithm with existing track, if match at
Function is then added into corresponding track；If it fails to match, predict this track in present frame by Kalman filtering algorithm
Present position；
(4) it if certain track continuous several times are matched to moving object not successfully, that is, continues through Kalman filtering algorithm and measures in advance
The number of the tracing point arrived is greater than particular value A, then it is assumed that this track terminates；
(5) if the number of the point of certain track is greater than threshold value B, then it is assumed that the track is effective, otherwise it is assumed that it is invalid.
9. a kind of throwing object in high sky detection method based on computer vision as claimed in claim 8, it is characterised in that: step 4)
In, the decision condition of throwing object in high sky is as follows:
(1) if what the value of the ordinate of movement objective orbit persistently increased；
(2) the number M of throwing object in high sky movement objective orbit point is greater than preset threshold value T, i.e. M > T；
(3) throwing object in high sky movement objective orbit terminates from the boundary of monitoring range；
(4) projectile motion is accelerated in beginning；
(5) in throwing object in high sky moving target dropping process, the difference numerical of ordinate will not be too small always.
10. a kind of throwing object in high sky detection method based on computer vision as claimed in claim 8, it is characterised in that: step
4) in, obtain belong to throwing object in high sky target dish out position the step of are as follows:
(1) it is directed to every track, all the points on track are fitted using least square method, are independent variable, y using frame number
For numerical value on direction as dependent variable, fitting obtains a function；
(2) according to the function, from first point of existing track toward rewinding, until the numerical difference in the direction y of continuous two points is small
In threshold value A, then it is assumed that reach throw point, current point is saved as and is dished out to a track.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811015244.1A CN109309811B (en) | 2018-08-31 | 2018-08-31 | High-altitude parabolic detection system and method based on computer vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811015244.1A CN109309811B (en) | 2018-08-31 | 2018-08-31 | High-altitude parabolic detection system and method based on computer vision |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109309811A true CN109309811A (en) | 2019-02-05 |
| CN109309811B CN109309811B (en) | 2021-02-19 |
Family
ID=65224317
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811015244.1A Active CN109309811B (en) | 2018-08-31 | 2018-08-31 | High-altitude parabolic detection system and method based on computer vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109309811B (en) |
Cited By (36)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110275042A (en) * | 2019-05-07 | 2019-09-24 | 深圳市零壹移动互联系统有限公司 | A kind of throwing object in high sky detection method based on computer vision and radio signal analysis |
| CN110738686A (en) * | 2019-10-12 | 2020-01-31 | 四川航天神坤科技有限公司 | Static and dynamic combined video man-vehicle detection method and system |
| CN111079663A (en) * | 2019-12-19 | 2020-04-28 | 深圳云天励飞技术有限公司 | High-altitude parabolic monitoring method and device, electronic equipment and storage medium |
| CN111260693A (en) * | 2020-01-20 | 2020-06-09 | 北京中科晶上科技股份有限公司 | Detection method of high-altitude object throwing |
| CN111382692A (en) * | 2020-03-05 | 2020-07-07 | 深学科技(杭州)有限公司 | High altitude parabolic detection method, device and computer readable storage medium |
| CN111401311A (en) * | 2020-04-09 | 2020-07-10 | 苏州海赛人工智能有限公司 | High-altitude parabolic recognition method based on image detection |
| CN111476163A (en) * | 2020-04-07 | 2020-07-31 | 浙江大华技术股份有限公司 | High-altitude parabolic monitoring method and device and computer storage medium |
| CN111488799A (en) * | 2020-03-13 | 2020-08-04 | 安徽小眯当家信息技术有限公司 | Falling object identification method and system based on image identification |
| CN111523397A (en) * | 2020-03-31 | 2020-08-11 | 深圳市奥拓电子股份有限公司 | Intelligent lamp pole visual identification device, method and system and electronic equipment |
| CN111539388A (en) * | 2020-06-22 | 2020-08-14 | 湖南联智科技股份有限公司 | High-altitude parabolic monitoring method |
| CN111553256A (en) * | 2020-04-26 | 2020-08-18 | 上海天诚比集科技有限公司 | High-altitude parabolic early warning identification method based on object track identification |
| CN111553274A (en) * | 2020-04-28 | 2020-08-18 | 青岛聚好联科技有限公司 | High-altitude parabolic detection method and device based on trajectory analysis |
| CN111553257A (en) * | 2020-04-26 | 2020-08-18 | 上海天诚比集科技有限公司 | High-altitude parabolic early warning method |
| CN111627049A (en) * | 2020-05-29 | 2020-09-04 | 北京中科晶上科技股份有限公司 | High-altitude parabola determination method and device, storage medium and processor |
| CN111654664A (en) * | 2020-05-08 | 2020-09-11 | 浙江大华技术股份有限公司 | High-altitude parabolic detection method and system, computer equipment and storage medium |
| CN111723654A (en) * | 2020-05-12 | 2020-09-29 | 中国电子系统技术有限公司 | High-altitude parabolic detection method and device based on background modeling, YOLOv3 and self-optimization |
| CN111753574A (en) * | 2019-03-26 | 2020-10-09 | 顺丰科技有限公司 | Throw area positioning method, device, equipment and storage medium |
| CN111898511A (en) * | 2020-07-23 | 2020-11-06 | 北京以萨技术股份有限公司 | High-altitude parabolic detection method, device and medium based on deep learning |
| CN111931599A (en) * | 2020-07-20 | 2020-11-13 | 浙江大华技术股份有限公司 | High altitude parabolic detection method, equipment and storage medium |
| CN112016414A (en) * | 2020-08-14 | 2020-12-01 | 熵康(深圳)科技有限公司 | Method and device for detecting high-altitude parabolic event and intelligent floor monitoring system |
| CN112213716A (en) * | 2020-09-21 | 2021-01-12 | 浙江大华技术股份有限公司 | Object positioning method and device and electronic equipment |
| CN112258573A (en) * | 2020-10-16 | 2021-01-22 | 腾讯科技（深圳）有限公司 | Method and device for acquiring throwing position, computer equipment and storage medium |
| CN112288991A (en) * | 2020-09-08 | 2021-01-29 | 江苏澳联科技开发有限公司 | Intelligent alarm system for falling object |
| CN112347915A (en) * | 2020-11-05 | 2021-02-09 | 重庆览辉信息技术有限公司 | High-altitude throwing object distinguishing system |
| CN112365524A (en) * | 2020-11-10 | 2021-02-12 | 宁波博登智能科技有限责任公司 | High-altitude parabolic real-time alarm system based on time sequence images |
| CN112418069A (en) * | 2020-11-19 | 2021-02-26 | 中科智云科技有限公司 | High-altitude parabolic detection method and device, computer equipment and storage medium |
| CN112488024A (en) * | 2020-12-08 | 2021-03-12 | 平安国际智慧城市科技股份有限公司 | High-altitude parabolic tracing method and device, computer equipment and storage medium |
| CN112489069A (en) * | 2020-12-12 | 2021-03-12 | 江西洪都航空工业集团有限责任公司 | Vehicle window parabolic detection method based on real-time video |
| CN112733690A (en) * | 2020-12-31 | 2021-04-30 | 北京易华录信息技术股份有限公司 | High-altitude parabolic detection method and device and electronic equipment |
| CN113052878A (en) * | 2021-03-30 | 2021-06-29 | 北京睿芯高通量科技有限公司 | Multi-path high-altitude parabolic detection method and system for edge equipment in security system |
| CN113255697A (en) * | 2021-06-01 | 2021-08-13 | 南京图菱视频科技有限公司 | High-precision high-altitude parabolic detection system and method under complex scene |
| CN113269046A (en) * | 2021-04-28 | 2021-08-17 | 深圳市海清视讯科技有限公司 | High-altitude falling object identification method and system |
| CN113409360A (en) * | 2021-06-29 | 2021-09-17 | 深圳市商汤科技有限公司 | High altitude parabolic detection method and device, equipment and computer storage medium |
| CN113808409A (en) * | 2020-06-17 | 2021-12-17 | 华为技术有限公司 | Road safety monitoring method, system and computer equipment |
| CN115086531A (en) * | 2022-07-08 | 2022-09-20 | 浙江上视科技有限公司 | Intelligent data intelligent visual platform for smart park |
| CN115937267A (en) * | 2023-03-03 | 2023-04-07 | 北京灵赋生物科技有限公司 | Target track tracking method based on multi-frame video |
-
2018
- 2018-08-31 CN CN201811015244.1A patent/CN109309811B/en active Active
Cited By (55)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111753574A (en) * | 2019-03-26 | 2020-10-09 | 顺丰科技有限公司 | Throw area positioning method, device, equipment and storage medium |
| CN110275042A (en) * | 2019-05-07 | 2019-09-24 | 深圳市零壹移动互联系统有限公司 | A kind of throwing object in high sky detection method based on computer vision and radio signal analysis |
| CN110275042B (en) * | 2019-05-07 | 2021-07-13 | 深圳市零壹移动互联系统有限公司 | High-altitude parabolic detection method based on computer vision and radio signal analysis |
| CN110738686A (en) * | 2019-10-12 | 2020-01-31 | 四川航天神坤科技有限公司 | Static and dynamic combined video man-vehicle detection method and system |
| CN110738686B (en) * | 2019-10-12 | 2022-12-02 | 四川航天神坤科技有限公司 | Static and dynamic combined video man-vehicle detection method and system |
| CN111079663A (en) * | 2019-12-19 | 2020-04-28 | 深圳云天励飞技术有限公司 | High-altitude parabolic monitoring method and device, electronic equipment and storage medium |
| WO2021120867A1 (en) * | 2019-12-19 | 2021-06-24 | 深圳云天励飞技术股份有限公司 | High toss act monitoring method and device, electronic device and storage medium |
| CN111079663B (en) * | 2019-12-19 | 2022-01-11 | 深圳云天励飞技术股份有限公司 | High-altitude parabolic monitoring method and device, electronic equipment and storage medium |
| CN111260693B (en) * | 2020-01-20 | 2023-07-28 | 北京中科晶上科技股份有限公司 | High-altitude parabolic detection method |
| CN111260693A (en) * | 2020-01-20 | 2020-06-09 | 北京中科晶上科技股份有限公司 | Detection method of high-altitude object throwing |
| CN111382692A (en) * | 2020-03-05 | 2020-07-07 | 深学科技(杭州)有限公司 | High altitude parabolic detection method, device and computer readable storage medium |
| CN111488799A (en) * | 2020-03-13 | 2020-08-04 | 安徽小眯当家信息技术有限公司 | Falling object identification method and system based on image identification |
| CN111523397A (en) * | 2020-03-31 | 2020-08-11 | 深圳市奥拓电子股份有限公司 | Intelligent lamp pole visual identification device, method and system and electronic equipment |
| CN111523397B (en) * | 2020-03-31 | 2023-11-24 | 深圳市奥拓电子股份有限公司 | Intelligent lamp post visual identification device, method and system and electronic equipment thereof |
| CN111476163B (en) * | 2020-04-07 | 2022-02-18 | 浙江大华技术股份有限公司 | High-altitude parabolic monitoring method and device and computer storage medium |
| CN111476163A (en) * | 2020-04-07 | 2020-07-31 | 浙江大华技术股份有限公司 | High-altitude parabolic monitoring method and device and computer storage medium |
| CN111401311A (en) * | 2020-04-09 | 2020-07-10 | 苏州海赛人工智能有限公司 | High-altitude parabolic recognition method based on image detection |
| CN111553257A (en) * | 2020-04-26 | 2020-08-18 | 上海天诚比集科技有限公司 | High-altitude parabolic early warning method |
| CN111553256A (en) * | 2020-04-26 | 2020-08-18 | 上海天诚比集科技有限公司 | High-altitude parabolic early warning identification method based on object track identification |
| CN111553274A (en) * | 2020-04-28 | 2020-08-18 | 青岛聚好联科技有限公司 | High-altitude parabolic detection method and device based on trajectory analysis |
| CN111654664B (en) * | 2020-05-08 | 2022-04-08 | 浙江大华技术股份有限公司 | High-altitude parabolic detection method and system, computer equipment and storage medium |
| CN111654664A (en) * | 2020-05-08 | 2020-09-11 | 浙江大华技术股份有限公司 | High-altitude parabolic detection method and system, computer equipment and storage medium |
| CN111723654A (en) * | 2020-05-12 | 2020-09-29 | 中国电子系统技术有限公司 | High-altitude parabolic detection method and device based on background modeling, YOLOv3 and self-optimization |
| CN111723654B (en) * | 2020-05-12 | 2023-04-07 | 中国电子系统技术有限公司 | High-altitude parabolic detection method and device based on background modeling, YOLOv3 and self-optimization |
| CN111627049B (en) * | 2020-05-29 | 2023-09-26 | 北京中科晶上科技股份有限公司 | Method and device for determining high-altitude parabolic object, storage medium and processor |
| CN111627049A (en) * | 2020-05-29 | 2020-09-04 | 北京中科晶上科技股份有限公司 | High-altitude parabola determination method and device, storage medium and processor |
| CN113808409A (en) * | 2020-06-17 | 2021-12-17 | 华为技术有限公司 | Road safety monitoring method, system and computer equipment |
| CN111539388A (en) * | 2020-06-22 | 2020-08-14 | 湖南联智科技股份有限公司 | High-altitude parabolic monitoring method |
| CN111539388B (en) * | 2020-06-22 | 2020-10-09 | 湖南联智科技股份有限公司 | High-altitude parabolic monitoring method |
| CN111931599B (en) * | 2020-07-20 | 2023-04-18 | 浙江大华技术股份有限公司 | High altitude parabolic detection method, equipment and storage medium |
| CN111931599A (en) * | 2020-07-20 | 2020-11-13 | 浙江大华技术股份有限公司 | High altitude parabolic detection method, equipment and storage medium |
| CN111898511A (en) * | 2020-07-23 | 2020-11-06 | 北京以萨技术股份有限公司 | High-altitude parabolic detection method, device and medium based on deep learning |
| CN112016414A (en) * | 2020-08-14 | 2020-12-01 | 熵康(深圳)科技有限公司 | Method and device for detecting high-altitude parabolic event and intelligent floor monitoring system |
| CN112288991A (en) * | 2020-09-08 | 2021-01-29 | 江苏澳联科技开发有限公司 | Intelligent alarm system for falling object |
| CN112213716A (en) * | 2020-09-21 | 2021-01-12 | 浙江大华技术股份有限公司 | Object positioning method and device and electronic equipment |
| WO2022078182A1 (en) * | 2020-10-16 | 2022-04-21 | 腾讯科技（深圳）有限公司 | Throwing position acquisition method and apparatus, computer device and storage medium |
| CN112258573B (en) * | 2020-10-16 | 2022-11-15 | 腾讯科技（深圳）有限公司 | Method and device for acquiring throwing position, computer equipment and storage medium |
| CN112258573A (en) * | 2020-10-16 | 2021-01-22 | 腾讯科技（深圳）有限公司 | Method and device for acquiring throwing position, computer equipment and storage medium |
| CN112347915B (en) * | 2020-11-05 | 2024-03-12 | 重庆览辉信息技术有限公司 | High-altitude throwing object distinguishing system |
| CN112347915A (en) * | 2020-11-05 | 2021-02-09 | 重庆览辉信息技术有限公司 | High-altitude throwing object distinguishing system |
| CN112365524B (en) * | 2020-11-10 | 2024-03-29 | 宁波博登智能科技有限公司 | High-altitude parabolic real-time alarm system based on time sequence image |
| CN112365524A (en) * | 2020-11-10 | 2021-02-12 | 宁波博登智能科技有限责任公司 | High-altitude parabolic real-time alarm system based on time sequence images |
| CN112418069A (en) * | 2020-11-19 | 2021-02-26 | 中科智云科技有限公司 | High-altitude parabolic detection method and device, computer equipment and storage medium |
| CN112488024A (en) * | 2020-12-08 | 2021-03-12 | 平安国际智慧城市科技股份有限公司 | High-altitude parabolic tracing method and device, computer equipment and storage medium |
| CN112489069A (en) * | 2020-12-12 | 2021-03-12 | 江西洪都航空工业集团有限责任公司 | Vehicle window parabolic detection method based on real-time video |
| CN112733690A (en) * | 2020-12-31 | 2021-04-30 | 北京易华录信息技术股份有限公司 | High-altitude parabolic detection method and device and electronic equipment |
| CN112733690B (en) * | 2020-12-31 | 2024-02-20 | 北京易华录信息技术股份有限公司 | High-altitude parabolic detection method and device and electronic equipment |
| CN113052878B (en) * | 2021-03-30 | 2024-01-02 | 北京中科通量科技有限公司 | Multipath high-altitude parabolic detection method and system for edge equipment in security system |
| CN113052878A (en) * | 2021-03-30 | 2021-06-29 | 北京睿芯高通量科技有限公司 | Multi-path high-altitude parabolic detection method and system for edge equipment in security system |
| CN113269046A (en) * | 2021-04-28 | 2021-08-17 | 深圳市海清视讯科技有限公司 | High-altitude falling object identification method and system |
| CN113255697A (en) * | 2021-06-01 | 2021-08-13 | 南京图菱视频科技有限公司 | High-precision high-altitude parabolic detection system and method under complex scene |
| CN113409360A (en) * | 2021-06-29 | 2021-09-17 | 深圳市商汤科技有限公司 | High altitude parabolic detection method and device, equipment and computer storage medium |
| CN115086531A (en) * | 2022-07-08 | 2022-09-20 | 浙江上视科技有限公司 | Intelligent data intelligent visual platform for smart park |
| CN115937267A (en) * | 2023-03-03 | 2023-04-07 | 北京灵赋生物科技有限公司 | Target track tracking method based on multi-frame video |
| CN115937267B (en) * | 2023-03-03 | 2023-10-24 | 北京灵赋生物科技有限公司 | Target track tracking method based on multi-frame video |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109309811B (en) | 2021-02-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109309811A (en) | A kind of throwing object in high sky detection system based on computer vision and method | |
| CN109872341A (en) | A kind of throwing object in high sky detection method based on computer vision and system | |
| Bhatti et al. | Weapon detection in real-time cctv videos using deep learning | |
| CN103902966B (en) | Video interactive affair analytical method and device based on sequence space-time cube feature | |
| US20160029031A1 (en) | Method for compressing a video and a system thereof | |
| CN103500324A (en) | Violent behavior recognition method based on video monitoring | |
| KR20210020723A (en) | Cctv camera device having assault detection function and method for detecting assault based on cctv image performed | |
| CN105894530A (en) | Detection and tracking solution scheme aiming at motion target in video | |
| Prakash et al. | Detecting and tracking of multiple moving objects for intelligent video surveillance systems | |
| CN108629935A (en) | A kind of method and system for climbing building pivot frame larceny based on video monitoring detection | |
| CN109299700A (en) | Subway group abnormality behavioral value method based on crowd density analysis | |
| Noor et al. | Object Detection: Harmful Weapons Detection using YOLOv4 | |
| Huszar et al. | Toward fast and accurate violence detection for automated video surveillance applications | |
| KR101492059B1 (en) | Real Time Object Tracking Method and System using the Mean-shift Algorithm | |
| Frejlichowski et al. | SmartMonitor: An approach to simple, intelligent and affordable visual surveillance system | |
| Huo et al. | 3DVSD: An end-to-end 3D convolutional object detection network for video smoke detection | |
| Satybaldina et al. | Development of an algorithm for abnormal human behavior detection in intelligent video surveillance system | |
| Shivthare et al. | Suspicious activity detection network for video surveillance using machine learning | |
| Shripriya et al. | Violence Detection System Using Resnet | |
| Chan | A robust target tracking algorithm for FLIR imagery | |
| CN106921846A (en) | Video mobile terminal legacy detection means | |
| Sarcar et al. | Detecting violent arm movements using cnn-lstm | |
| Wang | A pipeline algorithm for detection and tracking of pixel-sized target trajectories | |
| Wang et al. | Detection Anomaly in Video Based on Deep Support Vector Data Description. | |
| Alkanat et al. | Towards Scalable Abnormal Behavior Detection in Automated Surveillance |
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 |