CN109814074A - Multiple targets tracking based on image aspects processing - Google Patents
Multiple targets tracking based on image aspects processing Download PDFInfo
- Publication number
- CN109814074A CN109814074A CN201910057384.3A CN201910057384A CN109814074A CN 109814074 A CN109814074 A CN 109814074A CN 201910057384 A CN201910057384 A CN 201910057384A CN 109814074 A CN109814074 A CN 109814074A
- Authority
- CN
- China
- Prior art keywords
- multiple targets
- center
- coordinate
- sampling period
- 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.)
- Granted
Links
Abstract
The invention discloses a kind of multiple targets trackings based on image aspects processing, echo-signal including receiving within a sampling period to radar receiver is handled and carries out coordinate conversion and calculated, it establishes initial pictures matrix and calculates bianry image matrix, morphological image process is successively carried out to bianry image matrix and label handles and calculates the equivalent center vector of multiple targets in this sampling period, the original state at the track collection joint account multiple targets center of multiple targets is calculated using logical approach, the filter state for calculating incidence matrix and filter tracking track being adjusted according to it and predicts current filter pursuit path state as next sampling period multiple targets center, finally obtain multiple targets tracking result.The method of the present invention is used in the precision that radar tracking technical field has not only ensured multiple targets center calculation based on the multiple targets tracking that image aspects are handled, and more occurs to track multiple targets well under the deletion condition of part in multiple targets, has of great value practicability.
Description
Technical field
The invention belongs to radar target tracking technical field, in particular to it is a kind of based on image aspects processing multiple targets with
Track method.
Background technique
In modern war, the important means for having become military confrontation is carried out investigations and hit using unmanned air vehicle technique, by
The unmanned aerial vehicle group operation of unmanned plane composition is very huge for the threat of system of defense.And multiple targets can regard as one group each other
It is close close, the almost the same target of speed, the direction of motion in the short time.Therefore multiple targets center is tracked to maintain to group's mesh
The whole tracking of mark, does not consider each target in group, is tracked by the center to multiple targets, obtain each group in each frame
Position；There is important application value in target monitoring, unmanned plane multiple target formation motion tracking field in the sky for multiple targets tracking.
Since radar resolution cell is influenced by beam angle, the appearance of intensive multiple targets will affect detection echo precision so as to cause
The multiple targets observation that radar obtains is that part is distinguishable.Cause part distinguishable the reason is that due in the same group target with
The distance between target and angle have been more than radar resolution, and there is a situation where two target identifications often into a target,
The observation information of all single targets in group can not thus be obtained.This makes the center of multiple targets and the location error of real center
Increase, may cause when serious with losing multiple targets.
Summary of the invention
In order to solve the above-mentioned technical problem, the invention discloses a kind of multiple targets track sides based on image aspects processing
Method, the observation that this method is obtained after processing based on the echo in a radar receiver received sampling period, with figure
As morphology and connected component labeling method calculate the coordinate of the number of multiple targets and multiple targets equivalent center in this sampling period,
The track set at logical approach reasoning and calculation multiple targets center is further used, and determines the original state at multiple targets center, then
The multiple targets center in next sampling period is predicted according to the original state at multiple targets center, and next sampling is all
Multiple targets center vector in phase is associated with the multiple targets equivalent center position vector in corresponding next sampling period, most
Filter tracking parameter is adjusted according to association results afterwards and then multiple targets are tracked.This method is based on image aspects processing and improves
Precision of the radar plot preprocessing module to multiple targets center calculation, reduces the complexity of subsequent tracking, and make radar
Data processing module improves the efficiency to the tracking of multiple targets center when using Kalman filtering algorithm.
In order to achieve the above objectives, the present invention is realised by adopting the following technical scheme:
A kind of multiple targets tracking based on image aspects processing, described method includes following steps:
Step 1, the echo-signal received within a sampling period to radar receiver is handled, and is obtained radar and is existed
The distance and bearing angle that the total Q of target and each target measure in all echo-signals in this sampling period, and by straight
X-direction coordinate and Y of the distance and bearing angle of each target measurement in two-dimensional Cartesian coordinate system is calculated in angular coordinate conversion
Direction coordinate.
Step 2, initial pictures matrix I is established, and calculates bianry image matrix I '.
Step 3, morphological image process and label processing are successively carried out to bianry image matrix I ', obtain of multiple targets
Number L and image array Ipm, l=1,2 ..., a ..., L containing not isolabeling l.
Step 4, according to the image array Ipm containing not isolabeling l, calculate in this sampling period L multiple targets etc.
Imitate center vector Z, Z=(z1, z2... zl..., zL), wherein zl=(Xcl, Ycl), Xcl, YclIt respectively indicates in this sampling period
X-direction coordinate and Y-direction coordinate of the equivalent center of first of multiple targets in rectangular coordinate system.
Step 5, according to the equivalent center coordinate of L multiple targets in multiple sampling periods, using logical approach reasoning and calculation L
The track set H={ H of multiple targets1, H2..., Hl..., HL, wherein HlFor the stabilization track of first of multiple targets.
Step 6, the original state for calculating L multiple targets center, wherein the original state at first of multiple targets center are as follows:
Wherein, X_st(l)Indicate the original state of first of multiple targets, X_initial(l)、Y_initial(l)、It respectively indicates
X-direction coordinate and Y-direction coordinate of first of multiple targets center in original state, VX_initial(l)、VY_initial(l)Respectively
Indicate first of multiple targets center in the X-direction speed and Y-direction speed of original state.
Step 7, radar data processing module is on the basis of the original state at the L multiple targets center that step 6 obtains to L
Multiple targets start to track, if tracking total step number is K, initialize k=0.
Step 8, radar receiver executes step 1 to step after receiving all echo-signals in next sampling period
4, to obtain the equivalent center vector Z of F multiple targets in next sampling periodc, then Zc=(z1, z2... zf..., zF)。
Step 9, to the equivalent center vector Z of F multiple targets in next sampling periodcWith corresponding next sampling
The prediction center vector Z of L multiple targets in periodc' it is associated calculating, obtain incidence matrix C, next sampling
The prediction center vector Z of L multiple targets in periodc' be calculated in the sub-step 3 of the step 9；
The step 9 specifically includes following three sub-steps:
Sub-step 1 judges F and L relationship: if F > L, executing sub-step 2；If F=L, sub-step 3 is executed；
Sub-step 2 illustrates that segregation phenomenon occur in L multiple targets of tracking, then then step 4 continues to hold as F > L
Row step 5 to step 6 recalculates the original state at F multiple targets center, then executes the sub-step 3 in step 9 again；
Sub-step 3, enables k add 1, calculates the incidence matrix C.
Step 10, element c in the incidence matrix C obtained according to step 9flNumerical value to corresponding multiple targets filter tracking
Track is adjusted: being directed to cfl=1, indicate the equivalent center and first of multiple targets of f-th of multiple targets in next sampling period
Prediction core contextual, then the equivalent center coordinate of this f-th of multiple targets is used for the multiple targets in next sampling period
Track is filtered tracking, to obtain the filtering track of current F multiple targets；
Simultaneously by the filtering track of the current F multiple targets on the one hand as F group in new next sampling period
The filter state that the center of target is predicted, the on the other hand tracking result as F multiple targets.
Step 11, judge the value of k, if k < K, execute step 8 to 10, otherwise terminate to track.
Compared with prior art, a kind of multiple targets tracking based on image aspects processing provided by the invention utilizes figure
As the method for morphological process ensures that the offset at multiple targets observation center can be reduced in the case where multiple targets part is observed and being lost
Degree keeps the multiple targets equivalent observation center obtained more accurate, improves tracking stability；It is used in radar tracking technology neck
Domain improves radar plot preprocessing module to the precision of multiple targets center calculation, can handle the not high group's mesh of partial resolution
Mark, to improve the tracking accuracy at multiple targets center.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
It obtains other drawings based on these drawings.
Fig. 1 is a kind of process signal of multiple targets tracking based on image aspects processing provided in an embodiment of the present invention
Figure；
Fig. 2 is the schematic diagram before the multiple targets image procossing of the embodiment of the present invention；
Fig. 3 is the multiple targets image of the embodiment of the present invention through image aspects treated schematic diagram；
Fig. 4 is the multiple targets equivalent center pursuit path figure based on image aspects processing of the embodiment of the present invention；
Fig. 5 is the real trace figure of the multiple targets central observation of the embodiment of the present invention；
Fig. 6 is the range error schematic diagram of 1 centrode of multiple targets of the embodiment of the present invention；
Fig. 7 is the azimuth angle error schematic diagram of 1 centrode of multiple targets of the embodiment of the present invention；
Fig. 8 is the range error schematic diagram of 2 centrode of multiple targets of the embodiment of the present invention；
Fig. 9 is the azimuth angle error schematic diagram of 2 centrode of multiple targets of the embodiment of the present invention；
Figure 10 is the range error schematic diagram of 3 centrode of multiple targets of the embodiment of the present invention；
Figure 11 is the azimuth angle error schematic diagram of 3 centrode of multiple targets of the embodiment of the present invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments, is based on
The embodiment of the present invention, every other reality obtained by those of ordinary skill in the art without making creative efforts
Example is applied, shall fall within the protection scope of the present invention.
Fig. 1 show a kind of process of multiple targets tracking based on image aspects processing provided in an embodiment of the present invention
Schematic diagram.
As shown in Figure 1, a kind of multiple targets tracking based on image aspects processing provided in an embodiment of the present invention, including
Following steps:
Step 1, the echo-signal received within a sampling period to radar receiver is handled, and is obtained radar and is existed
The distance and bearing angle that the total Q of target and each target measure in all echo-signals in this sampling period, and by straight
X-direction coordinate and Y of the distance and bearing angle of each target measurement in two-dimensional Cartesian coordinate system is calculated in angular coordinate conversion
Direction coordinate, specific calculating process are as follows:
Firstly, successively carrying out pulse pressure, perseverance to all echo-signals that radar receiver receives within a sampling period
False-alarm detection and Monopulse estimation processing, to obtain mesh in all echo-signals that radar receives within this sampling period
The azimuth that the distance and each target that target sum Q, each target measure measure；
Secondly, the distance and bearing angle for according to the following formula measuring each target turns using rectangular coordinate conversion method
It is changed to each target and measures X-direction coordinate and Y-direction coordinate in two-dimensional Cartesian coordinate system:
xd=Rdcosθd, yd=Rdsinθd
Wherein, xd、ydIt respectively indicates d-th of target and measures X-direction coordinate and Y-direction seat in two-dimensional Cartesian coordinate system
Mark, RdIndicate the distance that d-th of target measures, θdIndicate the azimuth that d-th of target measures.
The step 1 by echo-signal that radar receiver receives within a sampling period carry out relevant treatment be in order to
Position of the total and each target for the target that actual observation arrives under polar coordinates is obtained, and the next to each target polar coordinates
It sets and carries out rectangular co-ordinate and be converted into the later period to lay the foundation with image procossing.
Step 2, initial pictures matrix I is established, and calculates bianry image matrix I ', specifically includes following sub-step:
Sub-step 1, if the range in targeted surveillance region is [x min, x max] × [y min, y max], wherein x
Min, x max respectively indicate the minimum value of targeted surveillance regional scope X-direction and the maximum value of X-direction, y min, y max difference
Indicate the minimum value of targeted surveillance regional scope Y-direction and the maximum value of Y-direction.
The pixel cell size of image is set as x_im × y_im simultaneously, wherein x_im, y_im respectively indicate image pixel list
The width and height of member.
The setting of the 1 targeted surveillance regional scope of sub-step be according in one sampling period of radar target it is total, every
What the coordinate of a target and repeatedly experimental summary determined.
Sub-step 2 calculates the dimension of the row and column of image array I, calculation formula are as follows:
Wherein, P indicates that the dimension of the row of image array I, N indicate the dimension of the column of image array I.
Sub-step 3 initializes all pixels unit in image array I, initial value 0.
Sub-step 4 calculates all targets in step 1 and measures the corresponding X-direction coordinate in image array I and Y-direction seat
Mark, calculation formula are as follows:
Wherein, m=1,2 ..., Q；Xm_Im、Ym_ Im respectively indicates m-th of target and measures the corresponding X in image array I
Direction coordinate and Y-direction coordinate；Xm_dkr、Ym_ dkr respectively indicates the X-direction coordinate and Y-direction coordinate of m-th of target measurement；<
> indicate to be rounded downwards；X min respectively indicates the minimum value of targeted surveillance regional scope X-direction, and y min respectively indicates target prison
Viewed area range Y-direction minimum value.
Sub-step 5 measures the corresponding X-direction coordinate in image array I according to obtained all targets and Y-direction is sat
Mark finds each target in image array I and measures a corresponding coordinate position, and by respective pixel at this coordinate position
The update of initial value 0 of unit is assigned a value of 1, to obtain the new image array I ' measured comprising all targets, then described image
Matrix I ' is bianry image matrix.
All targets that the step 2 obtains step 1 are established in image array I ' by the rectangular co-ordinate of each target,
In this way to be ready with morphological image process multiple targets tracking center and the key point of this method.
Step 3, morphological image process and label processing are successively carried out to bianry image matrix I ', obtain of multiple targets
Number L and image array Ipm, l=1,2 ..., a ..., L containing not isolabeling l, described image Morphological scale-space and mark
Remember the detailed process of processing are as follows:
Firstly, to bianry image matrix I ' carry out Morphological scale-space:
If structural element B is the matrix for the collar plate shape structure that radius is 8, according to image expansion algorithmTo binary map
As matrix I ' carry out expansive working, to obtain in the bianry image matrix Ip after border extended and the bianry image matrix Ip
Form multiple connected regions；In the bianry image matrix Ip dimension of row and column respectively with the middle row and column of image array I '
Dimension is consistent.
Secondly, calculating, institute is marked to multiple connected regions in bianry image matrix Ip using connected component labeling method
Stating label calculating process includes:
3a) to bianry image matrix Ip by row sequence progressively scan, if having in all pixels unit in the first row one or
Multiple pixel units are 1, then execute 3b) to 3d)；If all pixels unit is all 0 in the first row, second is continued to scan on
Row continues to scan on the third line until first appearing pixel list in the row of scanning if all pixels unit is all 0 in the second row
Member has 1, then executes 3b) to 3d)；
Bianry image matrix Ip 3b) is progressively scanned, is all 1 one sequence of composition contiguous pixels unit in every a line,
Record starting point, terminal and the line number at place of each sequence；
For row of the pixel unit for 1 is first appeared in all rows of bianry image matrix Ip, orderly to the institute in the row
Leie is marked, and corresponding label l is successively assigned a value of 1,2 ..., a '.
3c) by 3b) in the row based on the label number of all sequences label, to all sequences in next line successively into
Line flag, if s-th of sequence and 3b) sequence in the row has overlapping region, the label number of s-th of sequence mark
It is identical as the label number of this sequence mark in the 3b) row；If s-th of sequence and 3b) in the row two or two with
On sequence have overlapping region, then the label number and 3b of s-th of sequence mark) the two in the row or more than two sequences
The smallest label is number identical in the label number of label, and by the mark of the two or more than two sequence marks in the 3b) row
Mark is recorded as of equal value right；If s-th of sequence and 3b) all sequences in the row are all without overlapping region, s-th of sequence
The label number of column label is new label number, and as s=1, then new label l=a '+1, when s ≠ 1, then new label number is
It sorts according to the new label number that the next line sequence is marked；
And so on, all sequences in bianry image matrix Ip in all rows in back are successively marked.
The label number of the not homology equivalence pair recorded in 3c) 3d) is equivalent to the label of minimum value in the wherein number of label respectively
Number, complete to eliminate of equal value pair of calculating, all labels before then traversing and to all connected regions with natural number sequence weight
New label, to obtain the image array Ipm containing not isolabeling l, the maximum value of label is exactly the number L of connected domain, and
The number of connected domain is exactly the number of multiple targets, and the dimension of the row and column of described image matrix Ipm is respectively and in image array Ip
The dimension of row and column is consistent.
The step 3 is the most important thing of this method, is found in step 1 and is owned using morphological image and connected component labeling method
Population number, that is, multiple targets number of target and the geometric areas of each multiple targets, and the geometric areas mark of each multiple targets
Marked goes out to be presented in an image array Ipm.The center of the geometric areas of multiple targets each in this way is certain, even if
One or several targets missing in multiple targets nor affects on the center of the geometric areas of the multiple targets, to ensure that each group
The computational accuracy of target's center improves the tracking effect at multiple targets center.
Step 4, according to the image array Ipm containing not isolabeling l, calculate in this sampling period L multiple targets etc.
Imitate center vector Z, Z=(z1, z2... zl..., zL), wherein zl=(Xcl, Ycl), Xcl, YclIt respectively indicates in this sampling period
X-direction coordinate and Y-direction coordinate of the equivalent center of first of multiple targets in rectangular coordinate system:
It is corresponding by step 3 it is found that the connected region in image array Ipm is L, what radar receiver received
There are L multiple targets in all targets in one sampling period.
Firstly, calculating the coordinate at all connected region centers in image array Ipm, calculation formula is as follows:
Wherein,Indicate the X-direction coordinate at first of connected region center in image array Ipm,Indicate image array
The Y-direction coordinate at first of connected region center in Ipm；xln, yljRespectively indicate nth pixel unit in first of connected region
X-direction coordinate and j-th of pixel unit Y-direction coordinate；Lr, lc respectively indicate pixel unit in image array Ipm and get the bid
It is denoted as the X-direction pixel unit sum and Y-direction pixel unit sum of l；SlFor in first of connected region pixel unit it is total
Number.
Secondly, calculating the coordinate of the equivalent center of L multiple targets, calculation formula is as follows:
Wherein, Xcl, YclRespectively indicate X-direction coordinate and Y of the equivalent center of first of multiple targets in rectangular coordinate system
Direction coordinate.
Finally, according to the coordinate of the equivalent center of L multiple targets, obtain in this sampling period L multiple targets it is equivalent in
Heart vector Z=(z1, z2... zl..., zL), wherein zl=(Xcl, Ycl)。
The step 4 obtains the geometric areas center of each multiple targets using the method for seeking mathematics geometrical center, then
The geometric areas center of each multiple targets is converted in rectangular coordinate system to obtain the equivalent center coordinate of each multiple targets.
Step 5, according to the equivalent center coordinate of L multiple targets in multiple sampling periods, using logical approach reasoning and calculation L
The track set H={ H of multiple targets1, H2..., Hl..., HL, wherein HlFor the stabilization track of first of multiple targets.
One sampling period is a sampling time, then multiple sampling periods are multiple sampling times, if when radar sampling
Between be T, preferably the radar sampling time be the first four moment, then execute step 1 respectively as T=1, T=2, T=3 and T=4
To step 4, the equivalent center vector of L multiple targets is inscribed when obtaining corresponding: setting T=t, t=1,2,3,4, then t-th of moment L
The equivalent center vector of a multiple targets is Zt, Zt=(z(t, 1), z(t, 2)..., z(t, L)), wherein z(t, l)=(XcT, l, YcT, l),
XcT, l, YcT, lRespectively indicate X-direction coordinate and Y side of the equivalent center of t-th of moment, first of multiple targets in rectangular coordinate system
To coordinate.
According to an embodiment of the invention, the sampling time is preferably the first four moment to be according to logical approach reasoning and multiple
It is quasi-definite to test mould, the general sampling time is first three or four moment, can infer multiple multiple targets with logical approach
Stabilization track.
The equivalent center coordinate that L multiple targets are inscribed when to using first four, using L multiple targets of logical approach reasoning
Stablize track:
Firstly, the equivalent center of L multiple targets in first moment establishes initial related wave door with tachometric method, to falling into
The equivalent center of L multiple targets in second moment of initial correlation Bo Mennei all establishes potential track set.
Secondly, extrapolating to each subset track in above-mentioned potential track set, centered on extrapolation point, by track
Extrapolation error covariance determines wave Men great little at this time: if in the third moment presence fall into the multiple targets of related Bo Mennei etc.
The equivalent center of the multiple targets nearest from extrapolation point is then stored in corresponding potential track subset by effect center；If without group
The equivalent center of target falls into related Bo Mennei, then this possible subset track is deleted from potential track set.
Finally, continuing to extrapolate to remaining potential track, fall into related Bo Mennei's if existing in the 4th moment
Group's equivalent center will be then stored in corresponding potential track set and by this track from the nearest group's equivalent center of extrapolation point
It is determined as stablizing track；If falling into related Bo Mennei without group's equivalent center, this possibility is deleted from potential track set
Track.
After handling according to above-mentioned logical approach, stabilization track set H, the H={ H of L multiple targets can be obtained1, H2...,
Hl..., HL, wherein HlEquivalent center for first of multiple targets of stabilization track and first four moment of first of multiple targets is sat
The subset that cursor position is formed, then HlExpression formula be Hl={ h(l, 1)..., h(l, t)... h(l, 4), wherein element h(l, t)It is t-th
When inscribe the equivalent center coordinate position of first of multiple targets, i.e.,Xc(l, t)、Yc(l, t)It respectively indicates t-th
When inscribe X-direction coordinate and Y-direction coordinate of the equivalent center of first of multiple targets in rectangular coordinate system.
Step 6, the original state for calculating L multiple targets center, wherein the original state at first of multiple targets center are as follows:
Wherein, X_st(l)Indicate the original state at first of multiple targets center, X_initial(l)、Y_initial(l)Respectively
Indicate first of multiple targets center in the X-direction coordinate and Y-direction coordinate of original state, VX_iniital(l)、VY_initial(l)
First of multiple targets center is respectively indicated in the X-direction speed and Y-direction speed of original state.
The original state at first of multiple targets center is calculated according to the following formula:
Wherein, Xc(l, T)、Yc(l, T)The equivalent center of first of multiple targets is inscribed when respectively indicating the T in rectangular coordinate system
In X-direction coordinate and Y-direction coordinate, Xc(l, t+1)、Yc(l, t+1)The equivalent of first of multiple targets is inscribed when respectively indicating the t+1
X-direction coordinate and Y-direction coordinate of the center in rectangular coordinate system, Δ t were indicated between the time between continuous two sampling periods
Every.
The original state at the L multiple targets center that the step 6 is calculated is next sampling period i.e. the 5th moment
The center prediction of lower L multiple targets is laid the groundwork.
Step 7, radar data processing module is on the basis of the original state at the L multiple targets center that step 6 obtains to L
Multiple targets start to track, if tracking total step number is K, initialize k=0.
The sampling time that the step 7 starts L multiple targets tracking is the 5th moment, is inscribed when successively tracking next
L multiple targets.
Step 8, radar receiver executes step 1 to step after receiving all echo-signals in next sampling period
4, to obtain the equivalent center vector Z of F multiple targets in next sampling periodc, then Zc=(z1, z2... zf..., zF)。
According to an embodiment of the invention, radar receiver reception first to be tracked down before radar filters every secondary tracking
Then the echo-signal of all multiple targets in one sampling period executes all multiple targets that step 1 to step 4 is tracked
Equivalent center vector Zc, prepare for step 9 calculating.
Step 9, to the equivalent center vector Z of F multiple targets in next sampling periodcWith corresponding next sampling
The prediction center vector Z of L multiple targets in periodc' it is associated calculating, obtain incidence matrix C, next sampling
The prediction center vector Z of L multiple targets in periodc' be calculated in the sub-step 3 of the step 9；
The step 9 specifically includes following three sub-steps:
Sub-step 1 judges F and L relationship: if F > L, executing sub-step 2；If F=L, sub-step 3 is executed；
Sub-step 2 illustrates that segregation phenomenon occur in L multiple targets of tracking, then then step 4 continues to hold as F > L
Row step 5 to step 6 recalculates the original state at F multiple targets center, then executes the sub-step 3 in step 9 again；
Sub-step 3, enables k add 1, calculates the incidence matrix C.
According to an embodiment of the invention, first have to judge the relationship of F and L before calculating incidence matrix because it is practical with
Can there is a phenomenon where group target splittings when track: if F=L illustrate the multiple targets of tracking do not separate can continue it is subsequent
It calculates；If F > L illustrates that the multiple targets of tracking separate, this need to recalculate the equivalent center of F multiple targets and F
The original state of multiple targets is laid the groundwork for the continuation multiple targets tracking in later period.
Preferably, the specific calculating process of the sub-step 3 are as follows:
9a) judge the value of k, if k=1, successively execute sub-step 9b) and 9d), otherwise, successively execute sub-step 9c) and
9d)。
The step 9a) judge the value of k after, there are two kinds to execute calculating, when only multiple targets start tracking in practice
Use successively execution sub-step 9b) and 9d), this is because being tracked in this sampling period for the first time without the filtering of radar filter tracking
State, behind calculate when track be carried out sub-step 9c) and 9d), utilize the filter state of radar filter tracking to predict group
The center of target.
9b) use following formula in next sampling period on the basis of the original state at L in step 6 multiple targets center
The center of this L multiple targets is predicted, to obtain the prediction centre bit of L multiple targets in next sampling period
It sets:
X_pre(l)=X_initial(l)+VX_initial(l)·Δt
Y_pre(l)=Y_initial(l)+VY_initial(l)·Δt
Wherein, X_pre(l)、Y_pre(l)Respectively indicate the X of the pre- measured center of first of multiple targets in next sampling period
Direction coordinate and Y-direction coordinate, Δ t indicate the time interval between continuous two sampling periods；
The prediction center vector Z of L multiple targets in next sampling period then can be obtainedc', Zc'=(z1',
z2' ... zl' ..., zL'), wherein zl'=(X_pre(l), Y_pre(l))。
9c) use following formula to L group's mesh in next sampling period with the filter state at current L multiple targets center
Target center is predicted, to obtain the prediction center of first of multiple targets in next sampling period:
X_pre(l)=X_f(l)+VX_f(l)·Δt
Y_pre(l)=Y_f(l)+VY_f(l)·Δt
Wherein, X_pre(l)、Y_pre(l)Respectively indicate the X of the pre- measured center of first of multiple targets in next sampling period
Direction coordinate and Y-direction coordinate, X_f(l)、Y_f(l)Respectively indicate the filtered X-direction coordinate in current first of multiple targets center and
Y-direction coordinate, VX_f(l)、VY_f(l)Respectively indicate the filtered X-direction speed in current first of multiple targets center and Y-direction speed
Degree, Δ t indicate the sampling time interval of two continuous frames observation information；
The prediction center vector Z of L multiple targets in next sampling period then can be obtainedc', Zc'=(z1',
z2' ... zl' ..., zL'), wherein zl'=(X_pre(l), Y_pre(l))。
9d) to the equivalent center vector Z of F multiple targets in next sampling periodcWith corresponding next sampling week
The prediction center vector Z of L multiple targets in phasec' it is associated calculating:
If rectangular wave door threshold value Ux、Uy, the element c of incidence matrix C is determined according to the following formulafl:
Wherein, cflIndicate the element of f row l column in incidence matrix C, | | expression takes absolute value, and & is indicated and operation,
Xcf, YcfRespectively indicate X-direction coordinate of the equivalent center of f-th of multiple targets in next sampling period in rectangular coordinate system
With Y-direction coordinate.
Step 10, element c in the incidence matrix C obtained according to step 9flNumerical value to corresponding multiple targets filter tracking
Track is adjusted: being directed to cfl=1, indicate the equivalent center and first of multiple targets of f-th of multiple targets in next sampling period
Prediction core contextual, then the equivalent center coordinate of this f-th of multiple targets is used for the multiple targets in next sampling period
Track is filtered tracking, to obtain the filtering track of current F multiple targets；Simultaneously by the filter of the current F multiple targets
On the one hand filter state that wave track is predicted as the center of F multiple targets in new next sampling period, separately
On the one hand the tracking result as F multiple targets.
Step 11, judge the value of k, if k < K, execute step 8 to 10, otherwise terminate to track.
It is verified below by way of effect of the emulation experiment to the above method provided in an embodiment of the present invention:
Emulation experiment environment:
Experimental situation: Inter (R) Core (TM) i7-7700CPU@3.60HGz, 64 Windows operating systems and
MATLAB 2014a simulation software.
Experimental data:
The total Q=24 of target in the sampling period that radar receiver receives shares multiple targets L=3, Mei Gequn
In separately include 8 targets, and the actual observation target number that radar detection goes out is 3-4, and multiple targets 1,2,3 originate in respectively
[- 8000-2000], [- 80002000], near [- 6000-4000], range error σr=20m, angle error σθ=
0.1rad.Radar resolution ratio is λr=20, angular resolution λθ=0.01rad.
Experimental result:
1. as Fig. 2 and Fig. 3 respectively indicates signal of the multiple targets image of the embodiment of the present invention through image aspects before and after the processing
Figure, i.e. Fig. 2 and Fig. 3 be respectively before the processing of multiple targets image expansion with the multiple targets image after expansion process, by two images it is found that
The image pixel elements of the same multiple targets form connected region after treatment, so as to calculate in each connected region
The heart and using the center of each connected region as the equivalent center of corresponding multiple targets, and then avoid part observed image and lack
In the case where mistake the phenomenon that multiple targets off-centring, the influence for observing missing image is minimized, and has ensured multiple targets center
The precision of calculating, so that the tracking of multiple targets center is more stable.
2. such as the tracking multiple targets equivalent center track based on image aspects processing that Fig. 4 is the embodiment of the present invention, Fig. 5 is
The real trace of the multiple targets central observation of the embodiment of the present invention, You Liangtu is it is found that tracking group's mesh based on image aspects processing
It marks equivalent center track and the real trace of the multiple targets central observation is very identical, do not occur biggish offset, also do not occur
Lose with the case where, illustrate the embodiment of the present invention be based on morphological image process obtained from multiple targets equivalent center be reliable.
3. if Fig. 6 and Fig. 7 is respectively the range error and azimuth angle error of 1 centrode of multiple targets of the embodiment of the present invention
Schematic diagram, Fig. 8 and Fig. 9 are respectively the range error and azimuth angle error signal of 2 centrode of multiple targets of the embodiment of the present invention
Figure, Figure 10 and Figure 11 are respectively the range error and azimuth angle error schematic diagram of 3 centrode of multiple targets of the embodiment of the present invention,
As seen from the figure, the range error at 1 center of multiple targets is -20m~25m, and azimuth angle error is -0.3~0.5 degree of degree；In multiple targets 2
The range error of the heart is -25m~20m, and azimuth angle error is -0.5~0.5 degree of degree；The range error at 3 center of multiple targets is -20m
~40m, azimuth angle error are -0.5~0.3 degree of degree, illustrate that the embodiment of the present invention passes through multiple targets equivalent center set and group's mesh
Incidence matrix between mark prediction center makes filter tracking performance to adjust radar filter tracking parameter, no matter from distance
Or it still is able in angle all with good tracking effect, and in the case where deletion condition is observed in multiple targets part keep good
Tracking effect.
By emulation experiment it is found that the method for the present invention is used in radar tracking based on the multiple targets tracking that image aspects are handled
Technical field has not only ensured the precision of multiple targets center calculation, more can be in the case where part deletion condition occurs for multiple targets well
Multiple targets are tracked, there is of great value practicability.
Those of ordinary skill in the art will appreciate that: realize that all or part of the steps of above method embodiment can pass through
The relevant hardware of program instruction is completed, and program above-mentioned can store in computer-readable storage medium, which exists
When execution, step including the steps of the foregoing method embodiments is executed；And storage medium above-mentioned includes: ROM, RAM, magnetic or disk
Etc. the various media that can store program code.
The above description is merely a specific embodiment, but scope of protection of the present invention is not limited thereto, any
Those familiar with the art in the technical scope disclosed by the present invention, can easily think of the change or the replacement, and should all contain
Lid is within protection scope of the present invention.Therefore, protection scope of the present invention should be based on the protection scope of the described claims.
Claims (8)
1. a kind of multiple targets tracking based on image aspects processing, which is characterized in that the described method comprises the following steps:
Step 1, the echo-signal received within a sampling period to radar receiver is handled, and is obtained radar and is originally being adopted
The distance and bearing angle that the total Q of target and each target measure in all echo-signals in the sample period, and sat by right angle
X-direction coordinate and Y-direction of the distance and bearing angle of each target measurement in two-dimensional Cartesian coordinate system is calculated in mark conversion
Coordinate；
Step 2, initial pictures matrix I is established, and calculates bianry image matrix I '；
Step 3, morphological image process and label processing are successively carried out to bianry image matrix I ', obtain the number L of multiple targets
And image array Ipm, l=1,2 ..., a ..., L containing not isolabeling l；
Step 4, according to the image array Ipm containing not isolabeling l, calculate in this sampling period L multiple targets it is equivalent in
Heart vector Z, Z=(z1, z2... zl..., zL), wherein zl=(Xcl, Ycl), Xcl, YclIt respectively indicates in this sampling period first
X-direction coordinate and Y-direction coordinate of the equivalent center of multiple targets in rectangular coordinate system；
Step 5, according to the equivalent center coordinate of L multiple targets in multiple sampling periods, using L group's mesh of logical approach reasoning and calculation
Target track set H={ H1, H2..., Hl..., HL, wherein HlFor the stabilization track of first of multiple targets；
Step 6, the original state for calculating L multiple targets center, wherein the original state at first of multiple targets center are as follows:
Wherein, X_st(l)Indicate the original state at first of multiple targets center, X_initial(l)、Y_initial(l)It respectively indicates
X-direction coordinate and Y-direction coordinate of first of multiple targets center in original state, VX_initial(l)、VY_initial(l)Respectively
Indicate first of multiple targets center in the X-direction speed and Y-direction speed of original state；
Step 7, radar data processing module is on the basis of the original state at the L multiple targets center that step 6 obtains to L group's mesh
Mark starts to track, if tracking total step number is K, initializes k=0；
Step 8, radar receiver, which receives, executes step 1 after all echo-signals in next sampling period to step 4, from
And obtain the equivalent center vector Z of F multiple targets in next sampling periodc, then Zc=(z1, z2... zf..., zF)；
Step 9, to the equivalent center vector Z of F multiple targets in next sampling periodcWith corresponding next sampling period
The prediction center vector Z of interior L multiple targetsc' it is associated calculating, obtain incidence matrix C, next sampling period
The prediction center vector Z of interior L multiple targetsc' be calculated in the sub-step 3 of the step 9；
The step 9 specifically includes following three sub-steps:
Sub-step 1 judges F and L relationship: if F > L, executing sub-step 2；If F=L, sub-step 3 is executed；
Sub-step 2 illustrates that segregation phenomenon occur in L multiple targets of tracking, then then step 4 continues to execute step as F > L
Rapid 5 recalculate the original state at F multiple targets center to step 6, then execute the sub-step 3 in step 9 again；
Sub-step 3, enables k add 1, calculates the incidence matrix C；
Step 10, element c in the incidence matrix C obtained according to step 9flNumerical value to corresponding multiple targets filter tracking track
It is adjusted: for cfl=1, indicate the pre- of the equivalent center of f-th of multiple targets in next sampling period and first multiple targets
The equivalent center coordinate of this f-th of multiple targets, then is used for the track of the multiple targets in next sampling period by measured center association
It is filtered tracking, to obtain the filtering track of current F multiple targets；
Simultaneously by the filtering track of the current F multiple targets on the one hand as F multiple targets in new next sampling period
The filter state predicted of center, the on the other hand tracking result as F multiple targets；
Step 11, judge the value of k, if k < K, execute step 8 to 10, otherwise terminate to track.
2. a kind of multiple targets tracking based on image aspects processing according to claim 1, which is characterized in that step
1 specifically includes following sub-step:
Sub-step 1 successively carries out pulse pressure, perseverance to all echo-signals that radar receiver receives within a sampling period
False-alarm detection and Monopulse estimation processing, to obtain mesh in all echo-signals that radar receives within this sampling period
The azimuth that the distance and each target that target sum Q, each target measure measure；
Sub-step 2, using rectangular coordinate conversion method, the distance and bearing angle for according to the following formula measuring each target is converted
X-direction coordinate and Y-direction coordinate in two-dimensional Cartesian coordinate system are measured for each target:
xd=Rdcosθd
yd=Rdsinθd
Wherein, xd、ydRespectively indicate X-direction coordinate of the distance and bearing angle of d-th of target measurement in two-dimensional Cartesian coordinate system
With Y-direction coordinate, RdIndicate the distance that d-th of target measures, θdIndicate the azimuth that d-th of target measures.
3. a kind of multiple targets tracking based on image aspects processing according to claim 1, which is characterized in that step
2 specifically include following sub-step:
Sub-step 1, if the range in targeted surveillance region is [xmin, xmax] × [ymin, ymax], wherein xmin, xmax difference
Indicate the minimum value of targeted surveillance regional scope X-direction and the maximum value of X-direction, ymin, ymax respectively indicate targeted surveillance area
The minimum value of domain range Y-direction and the maximum value of Y-direction；
The pixel cell size of image is set as x_im × y_im simultaneously, wherein x_im, y_im respectively indicate image pixel elements
Width and height；
Sub-step 2 calculates the dimension of the row and column of image array I, calculation formula are as follows:
Wherein, P indicates that the dimension of the row of image array I, N indicate the dimension of the column of image array I；
Sub-step 3 initializes all pixels unit in image array I, initial value 0；
Sub-step 4 calculates all targets measure in step 1 X-direction coordinate and Y-direction coordinate and respectively corresponds in image array I
In X-direction coordinate and Y-direction coordinate, calculation formula it is as follows:
Wherein, m=1,2 ..., Q, Xm_Im、Ym_ Im respectively indicates m-th of target and measures the corresponding X-direction in image array I
Coordinate and Y-direction coordinate, Xm_dkr、Ym_ dkr respectively indicates the X-direction coordinate and Y-direction coordinate of the measurement of m-th of target,<>table
Show downward rounding, xmin indicates the minimum value of targeted surveillance regional scope X-direction, and ymin indicates the targeted surveillance regional scope side Y
To minimum value；
Sub-step 5 measures corresponding X-direction coordinate and Y-direction coordinate in image array I according to obtained all targets,
Each target is found in image array I and measures a corresponding coordinate position, and by respective pixel unit at this coordinate position
The update of initial value 0 be assigned a value of 1, to obtain the new image array I ' measured comprising all targets, then described image matrix
I ' is bianry image matrix.
4. a kind of multiple targets tracking based on image aspects processing according to claim 1, which is characterized in that step
3 specifically include:
Firstly, to bianry image matrix I ' carry out Morphological scale-space:
If structural element B is the matrix for the collar plate shape structure that radius is 8, according to image expansion algorithmTo bianry image square
Battle array I ' carry out expansive working, to obtain being formed in the bianry image matrix Ip after border extended and the bianry image matrix Ip
Multiple connected regions；The dimension of the row and column dimension with the middle row and column of image array I ' respectively in the bianry image matrix Ip
Unanimously；
Secondly, calculating, the mark is marked to multiple connected regions in bianry image matrix Ip using connected component labeling method
Remember that calculating process includes:
4a) bianry image matrix Ip is progressively scanned by row sequence, if having one or more in all pixels unit in the first row
Pixel unit is 1, then executes 4b) to 4d)；If all pixels unit is all 0 in the first row, the second row is continued to scan on, such as
All pixels unit is all 0 in the second row of fruit, then continues to scan on the third line until first appearing pixel unit in the row of scanning has
1, then execute 4b) to 4d)；
Bianry image matrix Ip 4b) is progressively scanned, is all 1 one sequence of composition contiguous pixels unit in every a line, record
Starting point, terminal and the line number at place of each sequence；
For first appeared in all rows of bianry image matrix Ip pixel unit be 1 row, to all sequences in the row according to
Secondary to be marked, corresponding label l is followed successively by l=1,2 ..., a；
4c) by 4b) in the row based on the label number of all sequences label, all sequences in next line are successively marked
Note, if a sequence in s-th sequence and the row has an overlapping region, the label number of s-th of sequence mark with it is described
The label number of this sequence mark is identical in row；If two or more sequences have weight in s-th of sequence and the row
Folded region, then the label number of s-th of sequence mark in the label number of the two or more than two sequence marks in the row
The smallest label is number identical, and the label number of the two in the row or more than two sequence marks is recorded as it is of equal value right；
If all sequences in s-th of sequence and the row are all without overlapping region, the label number of s-th of sequence mark is new
Label number, as s=1, then new label l=a+1, when s ≠ 1, then new label number is according to the next line sequence institute
The new label number sequence of label；
And so on, all sequences in bianry image matrix Ip in all rows in back are successively marked；
The label number of the not homology equivalence pair recorded in 4c) 4d) is equivalent to the label number of minimum value in the wherein number of label respectively, it is complete
At the calculating for eliminating of equal value pair, all labels before then traversing simultaneously are marked all connected regions with natural number sequence again
Note, to obtain the image array Ipm containing not isolabeling l, the maximum value of label is exactly the number L of connected domain, and is connected to
The number in domain is exactly the number of multiple targets, the dimension of the row and column of described image matrix Ipm respectively with row in image array Ip with
The dimension of column is consistent.
5. a kind of multiple targets tracking based on image aspects processing according to claim 1, which is characterized in that step
4 specifically include following sub-step:
Sub-step 1 calculates the coordinate at all connected region centers in image array Ipm according to the following formula:
Wherein,Indicate the X-direction coordinate at first of connected region center in image array Ipm,It indicates in image array Ipm
The Y-direction coordinate at first of connected region center, xln, yljRespectively indicate the side X of nth pixel unit in first of connected region
To coordinate and Y-direction coordinate, lr, lc respectively indicate the X-direction pixel unit for marking in pixel unit in image array Ipm and being
Sum and Y-direction pixel unit sum, SlFor the sum of pixel unit in first of connected region；
Sub-step 2 calculates the coordinate of the equivalent center of L multiple targets according to the following formula:
Wherein, Xcl, YclRespectively indicate X-direction coordinate and Y-direction of the equivalent center of first of multiple targets in rectangular coordinate system
Coordinate；
Sub-step 3 obtains the equivalent center of L multiple targets in this sampling period according to the coordinate of the equivalent center of L multiple targets
Vector Z=(z1, z2... zl..., zL), wherein zl=(Xcl, Ycl)。
6. a kind of multiple targets tracking based on image aspects processing according to claim 1, which is characterized in that step
5 detailed processes are as follows:
One sampling period is a sampling time, then multiple sampling periods are multiple sampling times, if the radar sampling time is
Preceding T moment, preferably radar sampling time are the first four moment, then execute step respectively as T=1, T=2, T=3 and T=4
Rapid 1 to step 4, and the equivalent center vector of L multiple targets is inscribed when obtaining corresponding: set T=t, t=1,2,3,4, then t-th when
The equivalent center vector for carving L multiple targets is Zt, Zt=(z(t, 1), z(t, 2)..., z(t, L)), wherein z(t, l)=(XcT, l, YcT, l),
XcT, l, YcT, lRespectively indicate X-direction coordinate and Y side of the equivalent center of t-th of moment, first of multiple targets in rectangular coordinate system
To coordinate；
The equivalent center coordinate that L multiple targets are inscribed when to using first four, using the stabilization of L multiple targets of logical approach reasoning
Track:
Firstly, the equivalent center of L multiple targets in first moment establishes initial related wave door with tachometric method, it is initial to falling into
The equivalent center of L multiple targets in second moment of related Bo Mennei all establishes potential track set；
Secondly, extrapolating to each subset track in above-mentioned potential track set, centered on extrapolation point, by course extrapolation
Error covariance determines wave Men great little at this time: if in the third moment presence fall into the multiple targets of related Bo Mennei it is equivalent in
The equivalent center of the multiple targets nearest from extrapolation point is then stored in corresponding potential track subset by the heart；If without multiple targets
Equivalent center fall into related Bo Mennei, then this possible subset track is deleted from potential track set；
Finally, continuing to extrapolate to remaining potential track, if there is the group etc. for falling into related Bo Mennei in the 4th moment
The group equivalent center nearest from extrapolation point is then stored in corresponding potential track set and determines this track by effect center
To stablize track；If falling into related Bo Mennei without group's equivalent center, this potential track is deleted from potential track set；
After handling according to above-mentioned logical approach, stabilization track set H, the H={ H of L multiple targets can be obtained1, H2..., Hl..., HL,
Wherein, HlFor the equivalent center coordinate position shape of first of multiple targets of stabilization track and first four moment of first of multiple targets
At subset, then HlExpression formula be Hl={ h(l, 1)..., h(l, t)... h(l, 4), wherein element h(l, t)Is inscribed when being t-th
The equivalent center coordinate position of l multiple targets, i.e.,Xc(l, t)、Yc(l, t)L is inscribed when respectively indicating t-th
X-direction coordinate and Y-direction coordinate of the equivalent center of a multiple targets in rectangular coordinate system.
7. a kind of multiple targets tracking based on image aspects processing according to claim 1, which is characterized in that step
6 specifically include:
The original state at first of multiple targets center is calculated according to the following formula:
Wherein, Xc(l, T)、Yc(l, T)X of the equivalent center of first of multiple targets in rectangular coordinate system is inscribed when respectively indicating the T
Direction coordinate and Y-direction coordinate, Xc(l, t+1)、Yc(l, t+1)The equivalent center of first of multiple targets is inscribed when respectively indicating the t+1
X-direction coordinate and Y-direction coordinate in rectangular coordinate system, Δ t indicate the time interval between continuous two sampling periods.
8. a kind of multiple targets tracking based on image aspects processing according to claim 1, which is characterized in that step
9 sub-step 3 specifically includes:
8a) judge the value of k, if k=1, successively execute sub-step 8b) and 8d), otherwise, successively execute sub-step 8c) and 8d)；
8b) on the basis of the original state at L in step 6 multiple targets center using following formula in next sampling period this
The center of L multiple targets is predicted, to obtain the prediction centre bit of first of multiple targets in next sampling period
It sets:
X_pre(l)=X_initial(l)+VX_initial(l)·Δt
Y_pre(l)=Y_initial(l)+VY_initial(l)·Δt
Wherein, X_pre(l)、Y_pre(l)Respectively indicate the X-direction of the pre- measured center of first of multiple targets in next sampling period
Coordinate and Y-direction coordinate, Δ t indicate the time interval between continuous two sampling periods；
The prediction center vector Z of L multiple targets in next sampling period then can be obtainedc', Zc'=(z1', z2' ...
zl' ..., zL'), wherein zl'=(X_pre(l), Y_pre(l))；
8c) with the filter state at current L multiple targets center using following formula to L multiple targets in next sampling period
Center is predicted, to obtain the prediction center of first of multiple targets in next sampling period:
X_pre(l)=X_f(l)+VX_f(l)·Δt
Y_pre(l)=Y_f(l)+VY_f(l)·Δt
Wherein, X_pre(l)、Y_pre(l)Respectively indicate the X-direction of the pre- measured center of first of multiple targets in next sampling period
Coordinate and Y-direction coordinate, X_f(l)、Y_f(l)Respectively indicate the filtered X-direction coordinate in current first of multiple targets center and the side Y
To coordinate, VX_f(l)、VY_f(l)The filtered X-direction speed in current first of multiple targets center and Y-direction speed are respectively indicated,
The sampling time interval of Δ t expression two continuous frames observation information；
The prediction center vector Z of L multiple targets in next sampling period then can be obtainedc', Zc'=(z1', z2' ...
zl' ..., zL'), wherein zl'=(X_pre(l), Y_pre(l))；
8d) to the equivalent center vector Z of F multiple targets in next sampling periodcWith L in corresponding next sampling period
The prediction center vector Z of a multiple targetsc' it is associated calculating:
If rectangular wave door threshold value Ux、Uy, the element c of incidence matrix C is determined according to the following formulafl:
Wherein, cflIndicate the element of f row l column in incidence matrix C, | | expression takes absolute value, and & is indicated and operation, Xcf, Ycf
Respectively indicate X-direction coordinate and Y-direction of the equivalent center of f-th of multiple targets in next sampling period in rectangular coordinate system
Coordinate.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910057384.3A CN109814074B (en) | 2019-01-22 | 2019-01-22 | Group target tracking method based on image morphological processing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910057384.3A CN109814074B (en) | 2019-01-22 | 2019-01-22 | Group target tracking method based on image morphological processing |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109814074A true CN109814074A (en) | 2019-05-28 |
| CN109814074B CN109814074B (en) | 2022-11-18 |
Family
ID=66603610
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910057384.3A Active CN109814074B (en) | 2019-01-22 | 2019-01-22 | Group target tracking method based on image morphological processing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109814074B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111208501A (en) * | 2019-12-30 | 2020-05-29 | 广州市番禺奥莱照明电器有限公司 | Multi-target aggregation method, device, equipment and storage medium based on radar detection |
| CN113030943A (en) * | 2021-03-05 | 2021-06-25 | 中国人民解放军空军工程大学航空机务士官学校 | Multi-target tracking algorithm for collecting azimuth range profile based on monopulse radar signals |
| CN113658220A (en) * | 2021-07-23 | 2021-11-16 | 哈尔滨工业大学 | Group target tracking method based on fusion of spatial clustering segmentation and multiple cost matrix association |
| CN114859300A (en) * | 2022-07-07 | 2022-08-05 | 中国人民解放军国防科技大学 | Radar radiation source data stream processing method based on graph connectivity |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004233136A (en) * | 2003-01-29 | 2004-08-19 | Mitsubishi Electric Corp | Target tracking device |
| CN108303692A (en) * | 2018-01-30 | 2018-07-20 | 哈尔滨工业大学 | A kind of multi-object tracking method of solution doppler ambiguity |
| CN109031279A (en) * | 2018-06-23 | 2018-12-18 | 中国人民解放军海军航空大学青岛校区 | Multiple multiple targets trackings with shape information |
-
2019
- 2019-01-22 CN CN201910057384.3A patent/CN109814074B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004233136A (en) * | 2003-01-29 | 2004-08-19 | Mitsubishi Electric Corp | Target tracking device |
| CN108303692A (en) * | 2018-01-30 | 2018-07-20 | 哈尔滨工业大学 | A kind of multi-object tracking method of solution doppler ambiguity |
| CN109031279A (en) * | 2018-06-23 | 2018-12-18 | 中国人民解放军海军航空大学青岛校区 | Multiple multiple targets trackings with shape information |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111208501A (en) * | 2019-12-30 | 2020-05-29 | 广州市番禺奥莱照明电器有限公司 | Multi-target aggregation method, device, equipment and storage medium based on radar detection |
| CN111208501B (en) * | 2019-12-30 | 2022-06-03 | 广州市番禺奥莱照明电器有限公司 | Multi-target aggregation method, device, equipment and storage medium based on radar detection |
| CN113030943A (en) * | 2021-03-05 | 2021-06-25 | 中国人民解放军空军工程大学航空机务士官学校 | Multi-target tracking algorithm for collecting azimuth range profile based on monopulse radar signals |
| CN113030943B (en) * | 2021-03-05 | 2023-08-18 | 中国人民解放军空军工程大学航空机务士官学校 | Multi-target tracking algorithm based on monopulse radar signal acquisition azimuth range profile |
| CN113658220A (en) * | 2021-07-23 | 2021-11-16 | 哈尔滨工业大学 | Group target tracking method based on fusion of spatial clustering segmentation and multiple cost matrix association |
| CN113658220B (en) * | 2021-07-23 | 2024-01-19 | 哈尔滨工业大学 | Group target tracking method based on fusion space clustering segmentation and multiple cost matrix association |
| CN114859300A (en) * | 2022-07-07 | 2022-08-05 | 中国人民解放军国防科技大学 | Radar radiation source data stream processing method based on graph connectivity |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109814074B (en) | 2022-11-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109814074A (en) | Multiple targets tracking based on image aspects processing | |
| CN104931966B (en) | A kind of spaceborne video SAR image processing methods based on DCS algorithms | |
| CN109100714A (en) | A kind of low slow small target tracking method based on polar coordinate system | |
| CN104730510B (en) | A kind of many radar track fusion methods | |
| CN108460341A (en) | Remote sensing image object detection method based on integrated depth convolutional network | |
| CN108535713A (en) | A kind of radar and the tracking of AIS joints and information fusion method | |
| CN104133983B (en) | Multiple-target Bernoulli distribution type fusion method based on weighting negative first-order RD (Renyi Divergence) | |
| CN108196240A (en) | Ground moving target track reconstruction method suitable for CSAR imaging | |
| CN111829964B (en) | Distributed remote sensing satellite system | |
| CN105652255B (en) | The spatial registration method of Radar Network System | |
| CN108205135A (en) | The radar video imaging method of quick rear orientation projection is merged based on no interpolation | |
| Karniely et al. | Sensor registration using neural networks | |
| CN105301584B (en) | The IPPHDF multiple-moving target tracking methods of fuzzy distance solution simultaneously | |
| CN112325883A (en) | Indoor positioning method for mobile robot with WiFi and visual multi-source integration | |
| CN116245916B (en) | Unmanned ship-oriented infrared ship target tracking method and device | |
| CN110501006A (en) | A kind of Dissimilar sensors joint track association and tracking | |
| CN108803655A (en) | A kind of UAV Flight Control platform and method for tracking target | |
| Zhao et al. | Imaging the Sun's far-side active regions by applying multiple measurement schemes on multiskip acoustic waves | |
| Fleck et al. | Robust tracking of reference trajectories for autonomous driving in intelligent roadside infrastructure | |
| CN108107427A (en) | Airborne/missile-borne array radar forword-looking imaging method based on super resolution technology | |
| Sun et al. | Vessel velocity estimation and tracking from Doppler echoes of T/RR composite compact HFSWR | |
| CN110751077B (en) | Optical remote sensing picture ship detection method based on component matching and distance constraint | |
| Chen et al. | Flying bird detection and hazard assessment for avian radar system | |
| Yan et al. | Moving targets detection for video SAR surveillance using multilevel attention network based on shallow feature module | |
| CN113589848B (en) | Multi-unmanned aerial vehicle detection, positioning and tracking system and method based on machine vision |
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 |