CN106846371A - A kind of birds wing vibration frequency measuring method and device based on video image - Google Patents
A kind of birds wing vibration frequency measuring method and device based on video image Download PDFInfo
- Publication number
- CN106846371A CN106846371A CN201611202815.3A CN201611202815A CN106846371A CN 106846371 A CN106846371 A CN 106846371A CN 201611202815 A CN201611202815 A CN 201611202815A CN 106846371 A CN106846371 A CN 106846371A
- Authority
- CN
- China
- Prior art keywords
- birds
- change
- moving target
- current frame
- image
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
Landscapes
- Catching Or Destruction (AREA)
- Image Analysis (AREA)
Abstract
The applicable field of computer technology of the present invention, there is provided a kind of birds wing vibration frequency measuring method and device based on video image, the method include:Two field picture is extracted from the Bird Flight video of input, birds moving target in the current frame image obtained to extraction is detected, to obtain shape and the position of birds moving target, birds moving target to detecting is tracked, the change in size signal changed over time with the shape for obtaining birds moving target, change in size signal is pre-processed, to obtain birds moving target, standardized change in size signal, according to standardized change in size signal, wing beating frequency to birds moving target is analyzed, so as to obtain the wing vibration frequency of birds in Bird Flight video, while the acquisition process of birds wing vibration frequency is simplified, ensure that birds wing pats the accuracy of identification of frequency.
Description
Technical field
Surveyed the invention belongs to field of computer technology, more particularly to a kind of birds wing vibration frequency based on video image
Amount method and device.
Background technology
Birds wing vibration frequency play the role of to its behavior it is important, wing vibration frequency anomaly can be used to sentence
The influences of the external environmental factor to birds such as disconnected chemicals and illumination.During Bird Flight wing vibration frequency determination for
It is very important to understand the state and offline mode of birds muscle.The frequency of each species wing vibration can be with other things
The frequency for planting wing vibration is overlapped, but the average frequency of each species wing vibration is significantly different with other species, therefore
Wing vibration frequency is not both the important information of identification different plant species.
Video can capture the information of all of free flight object, and the research vibrated to birds wing based on video is crucial
It is the information required for research can be extracted from video, although the research to birds wing vibration frequency has had for a long time
, but the method based on video research is also little.The most frequently used method for birds free flight detection is radar, the method
The changing features of temporal mode echo-signal are depended primarily on, this change is because birds cause body shape because of flight
Change.However, it was observed that signal intensity be mostly due to insect, bat or other flyers and weather
Caused by reason, therefore, to be subordinated in birds signal and distinguish birds signal is a problem for being relatively difficult to solve.In addition
A kind of method for birds free flight detection is analyzed with a kind of beating oscillator and with reference to the sense of hearing of the mankind, this
The method of kind contains the interference that birds are not attached to that part of flight during exercise, such as the respiratory movement of lung.
The content of the invention
It is an object of the invention to provide a kind of birds wing vibration frequency measuring method based on video image and device,
Aim to solve the problem that because prior art cannot provide a kind of effective birds wing vibration frequency measuring method, cause birds species to be known
Other accuracy rate is not high, be difficult to the problem of birds monitoring.
On the one hand, the invention provides a kind of birds wing vibration frequency measuring method based on video image, the side
Method comprises the steps:
Two field picture is extracted from the Bird Flight video of input, the birds in the current frame image obtained to extraction move mesh
Mark is detected, to obtain shape and the position of the birds moving target;
The birds moving target to detecting is tracked, to obtain the shape of the birds moving target with the time
The change in size signal of change;
The change in size signal is pre-processed, to obtain the birds moving target, standardized size change
Change signal;
According to the standardized change in size signal, the wing beating frequency to the birds moving target is divided
Analysis, so as to obtain the wing vibration frequency of birds in the Bird Flight video.
On the other hand, it is described the invention provides a kind of birds wing vibration frequency measuring device based on video image
Device includes:
Object detection unit, for extracting two field picture from the Bird Flight video of input, the present frame obtained to extraction
Birds moving target in image is detected, to obtain shape and the position of the birds moving target;
Target tracking unit, for being tracked to the birds moving target for detecting, is transported with obtaining the birds
The change in size signal that the shape of moving-target is changed over time;
Signal Pretreatment unit, for being pre-processed to the change in size signal, to obtain the birds motion mesh
Target, standardized change in size signal;And
Frequency analysis unit, for according to the standardized change in size signal, to the wing of the birds moving target
Wing is patted frequency and is analyzed, so as to obtain the wing vibration frequency of birds in the Bird Flight video.
The present invention extracts two field picture from the Bird Flight video of input, the birds in the current frame image obtained to extraction
Moving target is detected, to obtain shape and the position of birds moving target, the birds moving target to detecting carry out with
Track, the change in size signal changed over time with the shape for obtaining birds moving target, pre-processes to change in size signal,
To obtain birds moving target, standardized change in size signal, according to standardized change in size signal, birds are moved
The wing of target is patted frequency and is analyzed, and so as to obtain the wing vibration frequency of birds in Bird Flight video, is simplifying bird
While the acquisition process of class wing vibration frequency, it is ensured that birds wing pats the accuracy of identification of frequency.
Brief description of the drawings
Fig. 1 is the realization of the birds wing vibration frequency measuring method based on video image that the embodiment of the present invention one is provided
Flow chart;
Fig. 2 be the embodiment of the present invention two provide the birds wing vibration frequency measuring method based on video image in inspection
What the birds moving target for measuring was tracked realizes flow chart;
Fig. 3 is the structure of the birds wing vibration frequency measuring device based on video image that the embodiment of the present invention three is provided
Schematic diagram;
Fig. 4 is the preferred of the birds wing vibration frequency measuring device based on video image that the embodiment of the present invention three is provided
Structural representation;And
Fig. 5 is background in the birds wing vibration frequency measuring device based on video image that the embodiment of the present invention three is provided
The structural representation of updating block.
Specific embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, it is right below in conjunction with drawings and Examples
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.
Of the invention implementing is described in detail below in conjunction with specific embodiment:
Embodiment one:
Fig. 1 shows the birds wing vibration frequency measuring method based on video image that the embodiment of the present invention one is provided
Flow is realized, for convenience of description, the part related to the embodiment of the present invention is illustrate only, details are as follows:
In step S101, two field picture is extracted from the Bird Flight video of input, the current frame image obtained to extraction
In birds moving target detected, to obtain shape and the position of birds moving target.
In embodiments of the present invention, the aerial animal flight video in day is shot by setting camera, obtains the aerial birds in day
Moving image, it is preferable that camera is directed at sky and shoots birds sport video with 90 ° of elevations angle, so as to reduce other things in environment
Body influences on the detection band of birds moving target in image.
Cloud, weather such as change at the influence come to detection band in itself in considering sky, it is preferable that in working as being obtained to extraction
When birds moving target in prior image frame is detected, adjacent continuous of the current frame image relative to predetermined number are obtained first
The rate of change of two field picture, is updated according to rate of change to the background image of current frame image, and renewal is subtracted from current frame image
Background image afterwards, to obtain foreground image, binaryzation is carried out to foreground image, and thing is calculated from the foreground image after binaryzation
Body profile, to obtain one or more birds moving targets, so that accurately background image is separated from two field picture, and then
Obtain birds moving target.
It is further preferred that when the background image to current frame image is updated, using formula
Rate of change of the current frame image relative to the adjacent continuous two field picture of predetermined number is obtained, w, h are respectively the width of current frame image
Degree and height, Fi,jT () is the two field picture of t times and the pixel difference of Preset Time two field picture
Wherein, Ii,jThe pixel value at position (i, j) place in the two field picture of (t) expression t, Tf and γ is default value,
Ii,j(t- γ) is the pixel value at position (i, j) place in the two field picture at γ (i.e. (t- γ)) moment before moment t.
When rate of change v is located in the range of first, show that background image change is not very greatly, it is only necessary to current frame image
Background image carry out pixel renewal, so as to reduce the amount of calculation of context update.
Preferably, when rate of change is located in the range of first, the dynamic matrix that is associated with current frame image of acquisition, according to obtaining
The value of each element in the dynamic matrix for taking, determines the corresponding picture for carrying out pixel renewal in current frame image (t two field picture)
Vegetarian refreshments, and then use formulaTo the background image of current frame image
Pixel Bi,jT () is updated, wherein,It is weight parameter, Ii,jT () is the picture of the current frame image position (i, j) of t times
Element value.
Specifically, when rate of change is less than first threshold,Less than 0.5, when rate of change is equal to or more than the first threshold
Value and during less than Second Threshold,More than 0.5.The dynamic matrix associated with current frame imageWherein, λ is default value, when
Di,j(t) be 0 when, then it is assumed that change of the pixel (i, j) in continuous λ frames is smaller, now can determine that to background image position (i,
J) pixel value is updated.
When rate of change v is located in the range of second, background image is set to current frame image, i.e. background image Bi,j(t)
=Ii,j(t).Specifically, when rate of change is equal to or more than Second Threshold, it may be determined that change of the pixel (i, j) in continuous λ frames
It is larger, current frame image is directly now set into background image position.
In embodiments of the present invention, upgraded in time by background image, be effectively reduced background image change right
The influence of birds moving object detection, improves the detection efficiency of birds moving target.
In step s 102, the birds moving target for detecting is tracked, to obtain the shape of birds moving target
The change in size signal for changing over time.
In embodiments of the present invention, the birds moving target for detecting is tracked, to obtain birds moving target
The change in size signal that shape is changed over time.Preferred tracking mode refers to the description of embodiment two, will not be repeated here.
In step s 103, change in size signal is pre-processed, to obtain birds moving target, standardized chi
Very little variable signal.
In embodiments of the present invention, change with time signal obtaining bird moving target geomery, the change is believed
Breath is smoothed, average return-to-zero and standardization, so as to obtain standardized change in size signal.For side
Just subsequent descriptions, are referred to as original variable signal by the bird moving target geomery signal that changes with time herein.
Specifically, original variable signal F (n) for obtaining is filtered using standardized filter, after being smoothed
Variable signal GF (n), so as to provide basis subsequently to obtain accurate birds wing vibration frequency, wherein, F (n)=[f1,…,
fN], N is the length of signal.Afterwards, average return-to-zero is carried out to variable signal GF (n) after smoothing, in average return-to-zero
It is believed that two continuous and local minimum point Pi、Pi+1Between for wing pat a cycle.To the letter in each beating cycle
Number carry out average return-to-zero, signal AF (n) after average return-to-zero,
Wherein l=0 ..., L, L represent the number of local minimum.Finally, standard is carried out to signal AF (n) after average return-to-zero
Change is processed, and obtains standardized signal NF (n), wherein, NF (n)=AF (n)/Vmax, VmaxBe AF (n) (n=1 ..., N) most
Big value.
In step S104, according to standardized change in size signal, the wing beating frequency to birds moving target is entered
Row analysis, so as to obtain the wing vibration frequency of birds in Bird Flight video.
In embodiments of the present invention, frequency is patted to the wing of birds moving target by following steps to be analyzed, from
And obtain the wing vibration frequency of birds in Bird Flight video:
(1) Short Time Fourier Transform is carried out to standardized change in size signal, obtains representing multiple spectrum matrixs of wave spectrum.
In embodiments of the present invention, standardized change in size signal is divided by Short Time Fourier Transform (STFT)
Analysis, i.e., using formula To standard
The change in size signal of change is analyzed, wherein, x (m) is signal, and t is the time, and k represents frequency.W (m) is analysis window, and M is
The length of w (m), so, can combine the time-domain analysis of standardized change in size signal and frequency-domain analysis, for each
Time t value, can obtain spectrum matrix S (k) (k=0 ..., M-1), and one wave spectrum of each spectral moment matrix representation.
(2) the multiple spectrum matrixs obtained according to analysis calculate the average frequency of each wave spectrum.
In order to obtain quantitative local frequencies, the average frequency of each wave spectrum is calculatedThe average frequency is defined asWhereinWherein f (k) is k-th value of dot frequency, FsIt is sampling
Frequency.
(3) according to average frequency corresponding time point for being calculated, average oscillation frequency is ranked up, with to must put down
Equal frequency sequenceBy the wing vibration frequency that average frequency sequence is birds in Bird Flight video.
In embodiments of the present invention, examined by the birds moving target in the current frame image in birds video
Survey, to obtain shape and the position of birds moving target, the birds moving target to detecting is tracked, to obtain birds fortune
The change in size signal that the shape of moving-target is changed over time, pre-processes to change in size signal, to obtain birds motion
Target, standardized change in size signal, according to standardized change in size signal, the wing to birds moving target is patted
Frequency is analyzed, and so as to obtain the wing vibration frequency of birds in Bird Flight video, is simplifying birds wing vibration frequency
Acquisition process while, it is ensured that birds wing pat frequency accuracy of identification.
Embodiment two:
Fig. 2 is shown in the birds wing vibration frequency measuring method based on video image that the embodiment of the present invention one is provided
What the birds moving target to detecting was tracked realizes flow, for convenience of description, illustrate only and the embodiment of the present invention
Related part, details are as follows:
In step s 201, the Markov chain of birds moving target is initialized.
In embodiments of the present invention, it is every by what is detected for the ease of being tracked to the birds moving target for detecting
Individual birds moving target is expressed as Xi,t=(xi,t,yi,t,wi,t,hi,t,ri,t), it is also possible to referred to as birds moving target t
State, the Markov chain of t can randomly choose a sampling to be initialized from the Markov chain at t-1 moment.
Wherein (xi,t,yi,t) be birds moving target center, wi,tAnd hi,tIt is respectively width and length, the r of target areai,tIt is mesh
Mark relative to the orientation for shooting camera, Xi,tRepresent i-th birds moving target in t two field picture.
In step S202, the initial value of the Markov chain of definition and t according to birds moving target state is taken out
Sample produces the Markov Chain of t.
In embodiments of the present invention, when building the Markov Chain of birds moving target t, carried out by following manner
Build:
(1) by producing an equally distributed random number μ between [0,1], the value selection according to random number μ is corresponding
Sampling Modes are sampled, as a state in the Markov Chain of t;
In embodiments of the present invention, Sampling Modes include exchanging motor pattern and renewal motor pattern, if μ<0.5, choosing
Choose friends and change motor pattern and be sampled, otherwise selection updates motor pattern and is sampled.
(2) determine that t needs the birds moving target for updating;
In embodiments of the present invention, when it is determined that needing the birds moving target for updating, according to carrying for birds moving target
View select probability qv(i) (or qv(i, k)), determine a birds moving target i*(or two birds moving target i*,k*) enter
Row updates.
In motor pattern is updated,NtIt is the number of t birds moving target, its meaning is equiprobability choosing
Select a birds moving target i*, in the case where keeping other birds moving targets constant, update selected birds motion
Target i*State, that is, produce target i*State sampling.
In motor pattern is exchanged, the proposal select probability for defining birds moving target is
Wherein d (i, k) is the distance between target i, k.According to proposal select probability qv(i, k) selects two birds moving target i*、
k*, in the case where other birds moving target parameter constants are kept, exchange has selected birds moving target i*、k*Parameter.
(3) after it is determined that needing the birds moving target for updating, it is determined that needing the Ma Erke of the birds moving target of renewal
Husband's chain is constituted;
In embodiments of the present invention, so as to the Markov Chain after being updated.It is determined that needing the birds motion for updating
When the Markov Chain of target is constituted, current or new state vector Xi,tEach component in Xi,t-1On the basis of diffusion one symbol
The arbitrary width that average is zero Gaussian Profile is closed, renewal model is Xi,t/Xi,t-1=Xi,t-1+[Δx,Δy,Δw,Δh,
Δ r], wherein Δ x, Δ y, Δ w, Δ h and Δ r obey gaussian random migration, and the standard deviation of each Gauss migration determines phase
Answer the range of scatter of parameter.The change of birds shape can be timely covered for track window, Δ w and the corresponding Gausses of Δ h are set
The standard deviation sigma of random walkwAnd σhIt is 0.5.
(4) Sampling Modes for selecting are calculated with the receptance of sampling.
In embodiments of the present invention, the Sampling Modes for being selected according to step (1) calculate sample mode vector Xi,tReceiving
Rate.
In motor pattern is updated, receptance is defined as
Wherein,It is to update
To target i under motor pattern*The Markov Chain produced by more new model in step (3) is constituted or proposes that state is constituted, ZtIt is to work as
The observation information of preceding t,It is to stateObservation probability, φ (Xi,t,Xj,t) it is two interactions
Interactive potential energy between target i, j,It is and target i*The goal set for interacting, meanwhile, using byThe binary system that i.e. Kevin Smith (Kevin Smith) proposes is seen
Model is surveyed, precision p and recall rate r is defined to target i and j.
In motor pattern is exchanged, state acceptance rate is
(5) determined whether to receive current sample X according to the receptance for being calculated samplingi,t, with to birds moving target
Markov Chain is updated, to obtain the Markov Chain of t.
In embodiments of the present invention, if state acceptance rate a >=1 that obtains of computer, current sample is received, i.e.,If proposal is rejected, replicates previous sampling and sampled as current n-th, i.e.,
In step S203, according to the Markov Chain of the t obtained in step S202, tracked birds fortune is estimated
State vector of the moving-target in t.
In embodiments of the present invention, the application Markov chain Monte-Carlo (MCMC) in Bayes's multiple target tracking algorithm
Sampling, with produce N number of particle not weighted, i.e.,Maximum a posteriori probability side is utilized afterwards
Method estimates state vector of the tracked target in t.
In step S204, to the birds moving target that is obtained in step S203 in the state of t, birth and death processes is carried out
Detection, and change the number of tracked birds moving target.
In embodiments of the present invention, raw (birth) process detection refers to if detected previous in current frame image
Frame does not have matching target area, then it is assumed that be there is new target to enter or be blocked or with the target lost again
Occur.To the new target for entering, initialization tracking, and identity is distributed, to being blocked or the target with losing, from what is preserved
It is blocked or with the target information lost, identity when acquisition is initially tracked re-establishes tracking.Extremely (death) mistake
Journey detection refers to if for certain tracking box, not having matching region unit in the time range more than threshold value, then to recognize
For target disappears.Edge if in a two field picture disappears, it is believed that target leaves scene, deletes the tracking information of the target,
Otherwise it is assumed that target is blocked or with losing, the tracking information of target is preserved, and cancel the tracking to the target.
In step S205, the size signal of the birds moving target to being traced into step S204 is recorded.
In embodiments of the present invention, based on the tracking to birds moving target, record in each frame cover by each tracking box
The number of pixels of the binaryzation foreground image of lid.The counting of these number of pixels is believed as the size of birds moving target shape
Number, these size signals change with time and are changed with time signal as birds moving target size.
The embodiment of the present invention when the birds moving target to detecting is tracked, using in video extract connection figure
As detection birds moving target, and then the Markov Chain of birds moving target state is obtained, and then be accurately obtained birds fortune
Moving-target size changes with time signal, simplifies birds moving target size and changes with time the acquisition process of signal,
So as to provide more accurate data for the analysis of birds motion frequency.
One of ordinary skill in the art will appreciate that all or part of step in realizing above-described embodiment method can be
The hardware of correlation is instructed to complete by program, described program can be stored in a computer read/write memory medium,
Described storage medium, such as ROM/RAM, disk, CD.
Embodiment three:
Fig. 3 shows the birds wing vibration frequency measuring device based on video image that the embodiment of the present invention three is provided
Structure, for convenience of description, illustrate only the part related to the embodiment of the present invention, including:
Object detection unit 31, for extracting two field picture from the Bird Flight video of input, what extraction was obtained is current
Birds moving target in two field picture is detected, to obtain shape and the position of birds moving target;
Target tracking unit 32, for being tracked to the birds moving target for detecting, to obtain birds moving target
The change in size signal that changes over time of shape;
Signal Pretreatment unit 33, for being pre-processed to change in size signal, to obtain birds moving target, mark
The change in size signal of standardization;And
Frequency analysis unit 34, for according to standardized change in size signal, the wing to birds moving target to be patted
Frequency is analyzed, so as to obtain the wing vibration frequency of birds in Bird Flight video.
Preferably, as shown in figure 4, object detection unit 31 includes:
Context update unit 311, for obtaining change of the current frame image relative to the adjacent continuous two field picture of predetermined number
Rate, is updated according to rate of change to the background image of current frame image;And
Binarization unit 312 is right to obtain foreground image for subtracting the background image after updating from current frame image
Foreground image carries out binaryzation, and contour of object is calculated from the foreground image after binaryzation, to obtain birds moving target.
Wherein, as shown in figure 5, context update unit 311 may include
Rate of change acquiring unit 3111, for using formulaCurrent frame image is obtained relative to pre-
If the rate of change of the adjacent continuous two field picture of quantity, wherein, w, h are respectively the width and height of current frame image, Fi,jT () is t
The two field picture of time and the pixel difference of Preset Time two field picture;
Pixel updating block 3112, for when rate of change is located in the range of first, to the background image of current frame image
Carry out pixel renewal;And
Image update unit 3113, for when rate of change is located in the range of second, background image being set into present frame
Image.
In embodiments of the present invention, each unit of birds wing vibration frequency measuring device can be by corresponding hardware or software
Unit realizes that each unit can be independent soft and hardware unit, it is also possible to be integrated into a soft and hardware unit, be not used to herein
The limitation present invention.The specific embodiment of each unit refers to the description of embodiment one, no longer describes herein.
Presently preferred embodiments of the present invention is the foregoing is only, is not intended to limit the invention, it is all in essence of the invention
Any modification, equivalent and improvement made within god and principle etc., should be included within the scope of the present invention.
Claims (10)
1. a kind of birds wing vibration frequency measuring method based on video image, it is characterised in that methods described includes following
Step:
Two field picture is extracted from the Bird Flight video of input, the birds moving target in the current frame image obtained to extraction enters
Row detection, to obtain shape and the position of the birds moving target;
The birds moving target to detecting is tracked, and is changed over time with the shape for obtaining the birds moving target
Change in size signal;
The change in size signal is pre-processed, to obtain the birds moving target, standardized change in size letter
Number;
According to the standardized change in size signal, the wing beating frequency to the birds moving target is analyzed, from
And obtain the wing vibration frequency of birds in the Bird Flight video.
2. the method for claim 1, it is characterised in that the birds moving target in the current frame image obtained to extraction
The step of being detected, including:
Rate of change of the current frame image relative to the adjacent continuous two field picture of predetermined number is obtained, according to the rate of change pair
The background image of the current frame image is updated;
The background image after updating is subtracted from the current frame image, to obtain foreground image, two is carried out to the foreground image
Value, contour of object is calculated from the foreground image after the binaryzation, to obtain the birds moving target.
3. method as claimed in claim 2, it is characterised in that obtain the current frame image adjacent relative to predetermined number
The rate of change of sequential frame image, the step of being updated to the background image of the current frame image according to the rate of change, bag
Include:
Use formulaThe current frame image is obtained relative to the adjacent continuous two field picture of predetermined number
Rate of change, wherein, w, h are respectively the width and height of the current frame image, Fi,j(t) for the t times the two field picture with it is pre-
If the pixel difference of time two field picture;
When the rate of change is located in the range of first, the background image to the current frame image carries out pixel renewal;
When the rate of change is located in the range of second, the background image is set to the current frame image.
4. method as claimed in claim 3, it is characterised in that when the rate of change is located in the range of first, works as to described
The step of background image of prior image frame carries out pixel and updates, including:
When the rate of change is located in the range of first, the dynamic matrix that acquisition is associated with the current frame image;
The value of each element in dynamic matrix according to the acquisition, correspondence carries out pixel renewal in determining the current frame image
Pixel;
Use formulaTo the background image of the current frame image
Pixel Bi,jT () is updated, wherein,It is weight parameter, Ii,jT () is the current frame image position (i, j) of t times
Pixel value.
5. method as claimed in claim 2, it is characterised in that the step that the birds moving target to detecting is tracked
Suddenly, including:
Shape and position according to the birds moving target for acquiring, special karr algorithm is covered using Markov chain
The birds moving target is tracked, the birds moving target after the binaryzation is recorded in each two field picture
Number of pixels, the number of pixels is set to the size signal of the birds moving target;
According to each two field picture corresponding time point and the corresponding size signal, generate the birds moving target
The change in size signal that changes over time of shape.
6. the method for claim 1, it is characterised in that the step of being pre-processed to the change in size signal, bag
Include:
The change in size signal is smoothed, average return-to-zero and standardization, it is standardized to obtain
Change in size signal.
7. method as claimed in claim 6, it is characterised in that according to the standardized change in size signal, to the bird
The step of wing beating frequency of type games target is analyzed, including:
Short Time Fourier Transform is carried out to the standardized change in size signal, obtains representing multiple spectrum matrixs of wave spectrum;
The average frequency of each wave spectrum is calculated according to multiple spectrum matrixs;
According to the average frequency corresponding time point being calculated, the average oscillation frequency is ranked up, with to averagely
Frequency sequence, by the wing vibration frequency that the average frequency sequence is birds in the Bird Flight video.
8. a kind of birds wing vibration frequency measuring device based on video image, it is characterised in that described device includes:
Object detection unit, for extracting two field picture from the Bird Flight video of input, the current frame image obtained to extraction
In birds moving target detected, to obtain shape and the position of the birds moving target;
Target tracking unit, for being tracked to the birds moving target for detecting, to obtain the birds motion mesh
The change in size signal that target shape is changed over time;
Signal Pretreatment unit, for being pre-processed to the change in size signal, with obtain the birds moving target,
Standardized change in size signal;And
Frequency analysis unit, for according to the standardized change in size signal, the wing to the birds moving target to be clapped
Beat frequency to be analyzed, so as to obtain the wing vibration frequency of birds in the Bird Flight video.
9. device as claimed in claim 8, it is characterised in that the object detection unit includes:
Context update unit, for obtaining change of the current frame image relative to the adjacent continuous two field picture of predetermined number
Rate, is updated according to the rate of change to the background image of the current frame image;And
Binarization unit, for subtracting the background image after updating from the current frame image, to obtain foreground image, to described
Foreground image carries out binaryzation, and contour of object is calculated from the foreground image after the binaryzation, is moved with obtaining the birds
Target.
10. device as claimed in claim 9, it is characterised in that the context update unit includes:
Rate of change acquiring unit, for using formulaThe current frame image is obtained relative to present count
The rate of change of the adjacent continuous two field picture of amount, wherein, w, h are respectively the width and height of the current frame image, Fi,jT () is t
The two field picture of time and the pixel difference of Preset Time two field picture;
Pixel updating block, for when the rate of change is located in the range of first, to the background image of the current frame image
Carry out pixel renewal;And
Image update unit, for when the rate of change is located in the range of second, the background image being set into described working as
Prior image frame.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611202815.3A CN106846371A (en) | 2016-12-23 | 2016-12-23 | A kind of birds wing vibration frequency measuring method and device based on video image |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611202815.3A CN106846371A (en) | 2016-12-23 | 2016-12-23 | A kind of birds wing vibration frequency measuring method and device based on video image |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106846371A true CN106846371A (en) | 2017-06-13 |
Family
ID=59136958
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611202815.3A Pending CN106846371A (en) | 2016-12-23 | 2016-12-23 | A kind of birds wing vibration frequency measuring method and device based on video image |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106846371A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110659583A (en) * | 2019-04-26 | 2020-01-07 | 深圳市豪视智能科技有限公司 | Signal processing method and device and related product |
CN116630897A (en) * | 2023-07-20 | 2023-08-22 | 威海市三维工程测绘有限公司 | Airport bird repellent intelligent auxiliary control system based on image recognition |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103004719A (en) * | 2011-09-28 | 2013-04-03 | 沈金根 | High-efficiency mosquito eradication method and system |
CN104408406A (en) * | 2014-11-03 | 2015-03-11 | 安徽中科大国祯信息科技有限责任公司 | Staff off-post detection method based on frame difference method and background subtraction method |
CN105277929A (en) * | 2015-10-22 | 2016-01-27 | 北京理工大学 | Flapping frequency measurement method based on carrier phase doppler |
CN105590116A (en) * | 2015-12-18 | 2016-05-18 | 华南理工大学 | Bird image identification method based on head part alignment |
-
2016
- 2016-12-23 CN CN201611202815.3A patent/CN106846371A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103004719A (en) * | 2011-09-28 | 2013-04-03 | 沈金根 | High-efficiency mosquito eradication method and system |
CN104408406A (en) * | 2014-11-03 | 2015-03-11 | 安徽中科大国祯信息科技有限责任公司 | Staff off-post detection method based on frame difference method and background subtraction method |
CN105277929A (en) * | 2015-10-22 | 2016-01-27 | 北京理工大学 | Flapping frequency measurement method based on carrier phase doppler |
CN105590116A (en) * | 2015-12-18 | 2016-05-18 | 华南理工大学 | Bird image identification method based on head part alignment |
Non-Patent Citations (3)
Title |
---|
TAO YANG等: ""Real-time Multiple Objects Tracking with Occlusion Handing in Dynamic Scenes"", 《PROCEEDINGS OF THE 2005 IEEE COMPUTER SOCIETY CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION》 * |
XIAOYAN ZHANG等: ""Birds’ Flap Frequency Measure Based on Automatic Detection and Tracking in Captured Videos"", 《13TH INTERNATIONAL CONFERENCE ON BIOMEDICAL ENGINEERING, ICBME 2008》 * |
袁伟才等: ""一种有效的动态背景提取及更新方法"", 《计算机工程与应用》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110659583A (en) * | 2019-04-26 | 2020-01-07 | 深圳市豪视智能科技有限公司 | Signal processing method and device and related product |
CN110659583B (en) * | 2019-04-26 | 2023-11-28 | 深圳市豪视智能科技有限公司 | Signal processing method and device and related products |
CN116630897A (en) * | 2023-07-20 | 2023-08-22 | 威海市三维工程测绘有限公司 | Airport bird repellent intelligent auxiliary control system based on image recognition |
CN116630897B (en) * | 2023-07-20 | 2023-09-22 | 威海市三维工程测绘有限公司 | Airport bird repellent intelligent auxiliary control system based on image recognition |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107862705B (en) | Unmanned aerial vehicle small target detection method based on motion characteristics and deep learning characteristics | |
CN104899590B (en) | A kind of unmanned plane sensation target follower method and system | |
CN109785363A (en) | A kind of unmanned plane video motion Small object real-time detection and tracking | |
CN102494675B (en) | High-speed visual capturing method of moving target features | |
CN109448023B (en) | Satellite video small target real-time tracking method | |
CN105550678A (en) | Human body motion feature extraction method based on global remarkable edge area | |
CN106327526A (en) | Image object tracking method and image object tracking system | |
CN109859209B (en) | Remote sensing image segmentation method and device, storage medium and server | |
Jing et al. | AENN: A generative adversarial neural network for weather radar echo extrapolation | |
CN113808174B (en) | Radar small target tracking method based on full convolution network and Kalman filtering | |
CN109682378A (en) | A kind of unmanned plane indoor positioning and multi-object tracking method based entirely on visual information | |
CN102915545A (en) | OpenCV(open source computer vision library)-based video target tracking algorithm | |
CN103530893A (en) | Foreground detection method in camera shake scene based on background subtraction and motion information | |
Xiang et al. | Adaptive superpixel generation for SAR images with linear feature clustering and edge constraint | |
CN110111362A (en) | A kind of local feature block Similarity matching method for tracking target | |
CN110009060A (en) | A kind of robustness long-term follow method based on correlation filtering and target detection | |
Rizik et al. | Cost-efficient FMCW radar for multi-target classification in security gate monitoring | |
CN103942786B (en) | The self adaptation block objects detection method of unmanned plane visible ray and infrared image | |
Andriyanov et al. | Pattern recognition on radar images using augmentation | |
CN108765463A (en) | A kind of moving target detecting method calmodulin binding domain CaM extraction and improve textural characteristics | |
CN106846371A (en) | A kind of birds wing vibration frequency measuring method and device based on video image | |
CN113379789B (en) | Moving target tracking method in complex environment | |
CN115600101B (en) | Priori knowledge-based unmanned aerial vehicle signal intelligent detection method and apparatus | |
Wang et al. | Low-slow-small target tracking using relocalization module | |
CN116311067A (en) | Target comprehensive identification method, device and equipment based on high-dimensional characteristic map |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170613 |