CN110673132B - Real-time filtering method for trace point sequence for multi-frame joint detection and tracking - Google Patents

Real-time filtering method for trace point sequence for multi-frame joint detection and tracking Download PDF

Info

Publication number
CN110673132B
CN110673132B CN201910961671.7A CN201910961671A CN110673132B CN 110673132 B CN110673132 B CN 110673132B CN 201910961671 A CN201910961671 A CN 201910961671A CN 110673132 B CN110673132 B CN 110673132B
Authority
CN
China
Prior art keywords
trace
point
estimation
updated
frame
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.)
Active
Application number
CN201910961671.7A
Other languages
Chinese (zh)
Other versions
CN110673132A (en
Inventor
杨晓波
刘克柱
杨琪
张鹏辉
汤窈颖
李武军
易伟
孔令讲
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
University of Electronic Science and Technology of China
Original Assignee
University of Electronic Science and Technology of China
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by University of Electronic Science and Technology of China filed Critical University of Electronic Science and Technology of China
Priority to CN201910961671.7A priority Critical patent/CN110673132B/en
Publication of CN110673132A publication Critical patent/CN110673132A/en
Application granted granted Critical
Publication of CN110673132B publication Critical patent/CN110673132B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
    • G01S13/66—Radar-tracking systems; Analogous systems
    • G01S13/72—Radar-tracking systems; Analogous systems for two-dimensional tracking, e.g. combination of angle and range tracking, track-while-scan radar
    • G01S13/723—Radar-tracking systems; Analogous systems for two-dimensional tracking, e.g. combination of angle and range tracking, track-while-scan radar by using numerical data
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
    • G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
    • G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/14—Fourier, Walsh or analogous domain transformations, e.g. Laplace, Hilbert, Karhunen-Loeve, transforms

Abstract

The invention discloses a real-time filtering method for outputting a trace point sequence by a multi-frame joint detection and tracking algorithm, which is applied to the technical field of radar target detection and tracking and aims to solve the problems that the existing processing method is lack of real-time property for improving the target tracking precision, and the correlation between the trace point sequences output by the tracking algorithm before detection is difficult to calculate, the invention firstly obtains the trace point sets belonging to the same target at the same time according to the trace point sequences output by the tracking algorithm before batch processing, then, respectively performing Kalman filtering on the traces in the trace set to obtain corresponding updated state estimation and posterior covariance estimation, solving the minimum value of the updated posterior covariance estimation based on a determinant minimization criterion to obtain corresponding weight values, and finally obtaining the state estimation and the posterior covariance estimation after weighted summation by using the obtained weight values to complete filtering on a trace sequence; the method of the invention obviously improves the tracking precision of the tracking algorithm before detection.

Description

Real-time filtering method for trace point sequence for multi-frame joint detection and tracking
Technical Field
The invention belongs to the technical field of radar target detection and tracking, and particularly relates to a filtering technology for a trace point sequence output by a tracking algorithm before batch detection.
Background
For the data plane obtained by radar scanning, in order to track a target, a traditional tracking algorithm can perform constant false alarm detection on a single-frame data plane, the target is easily lost in the constant false alarm detection process, the detection and tracking performance of the radar on the weak target is seriously reduced, and the power push-away range of a radar system is weakened.
The pre-detection tracking technology is a technology capable of detecting and tracking a weak target signal. The biggest difference from the traditional detection method is that the tracking technology before detection does not perform threshold decision detection on data in a single frame, but performs joint accumulation on multi-frame data in a target echo. By increasing the dimension of time, multi-frame data are compared in a combined mode, real target echoes are separated by utilizing the difference between targets and clutter and noise, target loss caused by limited processing information of single-frame echo data detection is effectively avoided, and therefore the method can be used for detecting weak target signals. Common weak target pre-detection tracking algorithms include a dynamic programming-based pre-detection tracking algorithm, a maximum likelihood probability data fusion algorithm, a Hough transform pre-detection tracking algorithm, a particle filter-based pre-detection tracking algorithm, a random set theory-based pre-detection tracking algorithm and the like. However, the output of the pre-detection tracking algorithm is a plurality of trace point sequences with the same length, and the estimation of directly taking the trace point sequences as the target track leads to low target tracking precision. In the literature, "A novel dynamic programming algorithm for track-before-detect in radar systems, IEEE Transactions on Signal Processing, vol.61, No.10, pp.2608-2619,2013" proposes to smooth the trace point sequence output by the tracking algorithm before detection so as to improve the target tracking precision; but the details of the smoothing of the trace point sequence are not described. The document "Particle filtering for target tracking using spot-sequences of multi-frame track before detect," in 2015 IEEE Radar Conference (RadarCon), "IEEE 2015, pp.0495-0500", and its corresponding patent "a Particle filtering method for a multi-frame pre-detection tracking trace sequence, publication CN 104237853A" give a filtering method for a trace sequence, but this method focuses on the iteration problem between adjacent batches, while the present method focuses more on the fusion problem within the trace set. The document "a tracking adaptive for low-dependent target using poly-sequences of multi-frame detection, in 201619 th International Conference on Information Fusion (Fusion), IEEE,2016, pp.1427-1433" provides a smoothing method for a trace sequence output by a tracking algorithm before detection, but the method lacks rationality verification for an assumption of a measurement noise model, and a patent corresponding to the document "a smoothing filtering method for a trace sequence of multi-frame joint detection, the method proposed by publication No. CN 106226750A" is a smoothing algorithm, and only can improve target tracking accuracy before the current time, and lacks an effective method for improving target tracking accuracy at the current time; compared with a smoothing algorithm, the method has a real-time filtering tracking effect and can improve the target tracking precision at the current moment.
Disclosure of Invention
Aiming at the problems that the existing processing method lacks real-time performance for improving target tracking precision and correlation between trace point sequences output by a tracking algorithm before detection is difficult to calculate, the invention provides a real-time filtering method for outputting trace point sequences by a multi-frame joint detection and tracking algorithm, and the target tracking precision of the tracking algorithm before detection is improved.
The technical scheme adopted by the invention is as follows: for the convenience of describing the contents of the present invention, the following terms are first explained:
the term 1: trace point sequence
And (3) carrying out joint processing on the N frame data planes by a tracking algorithm before detection, and outputting a plurality of short track sections with the lengths of N.
The term 2: set of traces of dots
For a continuous batch pre-detection tracking algorithm, state estimates of the same target at the same time exist in a trace point sequence output by the continuous batch pre-detection tracking algorithm, and the repeated state estimates are collected in the same set and are called a trace point set.
The term 3: correlation between trace point sequences
The state estimates within the trace set are estimates of the same target state and thus there is a correlation between them.
The invention provides a real-time filtering algorithm for a multi-frame joint detection and tracking algorithm output trace point sequence, which comprises the following specific steps:
step 1, initializing system parameters
Make the total observed data plane frame numberL, the dimension of an observed data plane is A, and the combined processing frame number of a tracking algorithm before detection is N; the detection probabilities of the N frames are respectively
Figure GDA0003213770280000021
Each frame data plane sampling interval is TsA target state transition matrix F, an observation matrix H, a process noise covariance Q, an elliptic wave gate threshold gamma, and the current time k being N;
step 2, calculating the observation noise covariance of the tracking algorithm before detection
Respectively calculating the observation noise covariance matrix corresponding to each frame in the N frame data of the combined processing
Figure GDA0003213770280000022
N, wherein
Figure GDA0003213770280000023
Is the product of Kronecker, IΑRepresenting an identity matrix.
Step 3, point track set and track initialization
At the time k ═ N', the echo data of the 1 st, 2., N frames are read from the radar receiver, and the pre-detection tracking algorithm outputs a trace sequence
Figure GDA0003213770280000024
Thus obtaining the trace point set from all frames with the frame number of 1 being less than or equal to N being less than or equal to N at the moment of N
Figure GDA0003213770280000025
Wherein
Figure GDA0003213770280000031
The trace point sets are all under batch processing at the time of N' and come from the nth frame of the same target; and is initialized thereby
Figure GDA0003213770280000032
And Pn|n(1≤n≤k)。
Step 4, updating the trace point set
At time k +1, the slave radar is connectedReading echo data of a k-N +2, a, k +1 frame in a receiver, and outputting a trace point sequence by a tracking algorithm before detection
Figure GDA0003213770280000033
Thereby obtaining an updated trace point set
Figure GDA0003213770280000034
k-N +2 is not less than N is not less than k +1, and
Figure GDA0003213770280000035
step 5, for N ═ k-N +2
Figure GDA0003213770280000036
And (3) carrying out filtering treatment:
step 5.1, firstly, a trace point set of N-k-N +2 frames
Figure GDA0003213770280000037
Inner part
Figure GDA0003213770280000038
Spot trace
Figure GDA0003213770280000039
Performing Kalman filtering to obtain corresponding updated state estimation
Figure GDA00032137702800000310
Sum covariance estimation
Figure GDA00032137702800000311
Step 5.1.1 Aligning
Figure GDA00032137702800000312
Performing Kalman filtering to obtain corresponding updated state estimation
Figure GDA00032137702800000313
Sum covariance estimation
Figure GDA00032137702800000314
Including the following calculations:
intermediate variables
Figure GDA00032137702800000315
Intermediate variables
Figure GDA00032137702800000316
Wherein (·)TFor matrix transposition
Intermediate variables
Figure GDA00032137702800000317
Intermediate variables
Figure GDA00032137702800000318
Updated covariance estimation
Figure GDA00032137702800000319
Step 5.1.2, judging whether the requirement is met
Figure GDA00032137702800000320
If so, then there is an updated state estimate
Figure GDA00032137702800000321
Otherwise, there is
Figure GDA00032137702800000322
Step 5.1.3, judging whether the requirement is met
Figure GDA00032137702800000323
If yes, go to step 5.2; otherwise let t ═ t +1 and go to step 5.1.1.
Step 5.2, estimating all update states at the moment when N is k-N +2 according to determinant minimum covariance criterion
Figure GDA00032137702800000324
Sum covariance estimation
Figure GDA00032137702800000325
Solving the optimal combination:
Figure GDA0003213770280000041
Figure GDA0003213770280000042
obtaining corresponding weight vector
Figure GDA0003213770280000043
And solving to obtain the fused updated state estimation
Figure GDA0003213770280000044
Sum covariance estimation Pn|n：
Figure GDA0003213770280000045
Figure GDA0003213770280000046
Step 5.3, judging whether n is equal to k +1 or not, if yes, turning to step 6; otherwise, let n be n +1 and go to step 5.1.
Step 6, judging whether k +1 is equal to L or not, and if so, finishing updating all target states in the total observation time L frame; otherwise, let k be k +1, go to step 4.
The invention has the beneficial effects that: firstly, acquiring a trace point set belonging to the same target at the same moment according to a trace point sequence output by a tracking algorithm before batch processing detection, then respectively performing Kalman filtering on trace points in the trace point set to obtain corresponding updated state estimation and posterior covariance estimation, solving the minimum value of the updated posterior covariance estimation based on a determinant minimization criterion to obtain a corresponding weight, and finally obtaining the state estimation and the posterior covariance estimation after weighted summation by using the obtained weight; the method effectively solves the problem of complex correlation between trace sequences, realizes filtering of the trace sequences by utilizing the correlation between the trace sequences, and obviously improves the tracking precision of a tracking algorithm before detection; and has the following advantages:
1. by using a covariance cross algorithm, the direct calculation of the correlation between trace point sequences is ingeniously avoided;
2. filtering is carried out by combining a plurality of trace point sequences with overlapped trace points, and the correlation information among all trace point sequences output by a tracking algorithm before detection is fully utilized, so that the target tracking precision is improved in real time;
3. the solving process is simple, and the calculation complexity of the tracking algorithm before detection is reduced.
Drawings
FIG. 1 is a general flow diagram of the present invention.
FIG. 2 is a schematic diagram of an algorithm.
FIG. 3 is a graph comparing the effect of the solution of the present invention and the conventional method;
wherein, fig. 3(a) is a comparison graph of the real target trajectory and the estimated trajectory of the algorithm in a single scene (SNR ═ 5 dB); FIG. 3(b) is a graph comparing the RMES curves for KF-CI-DP-TBD (the algorithm of the present invention) and conventional DP-TBD (the conventional processing method) with an SNR of 6 dB; FIG. 3(c) is a graph comparing the RMES curves for KF-CI-MFDT and MFDT at SNR of 10 dB.
Detailed Description
In order to facilitate the understanding of the technical contents of the present invention by those skilled in the art, the following further explains the technical contents of the present invention with reference to fig. 1 to 3.
As shown in fig. 1, which is a general flow chart of the present invention, the implementation process includes steps 1-6, and all the steps and conclusions are verified to be correct on Matlab2017 b.
Step 1: the initialization of the system parameters is carried out,
specific simulation parameters are given in the embodiment: let the total number of frames of the observation data plane be L equal to 20, the dimension of the observation data plane be A equal to 2, and be largeIs as small as mx×ny70 × 70, in cells; the joint processing frame number of the tracking algorithm before detection is N-6; the detection probabilities of 6 frames are respectively PD＝[0.8058,0.8922,0.9024,0.8901,0.8832,0.8287]Each frame data plane sampling interval is T s1s, 16 elliptic wave gate threshold gamma, 6 current time k, a target state transition matrix F, an observation matrix H and a process noise covariance Q; the number of monte carlo trials was 500.
Figure GDA0003213770280000051
Step 2, calculating the observation noise covariance of the tracking algorithm before detection
Respectively calculating the observation noise covariance matrix corresponding to each frame in the N frame data of the combined processing
Figure GDA0003213770280000052
N, wherein
Figure GDA0003213770280000053
Is the product of Kronecker, I2＝[1,0；0,1]。
Figure GDA0003213770280000054
Representing the target detection probability of the mth frame in the N frame data processed in a combined mode;
step 3, point track set and track initialization
At the time k ═ N', the echo data of the 1 st, 2., N frames are read from the radar receiver, and the pre-detection tracking algorithm outputs a trace sequence
Figure GDA0003213770280000055
Thus obtaining the trace point set from all frames with the frame number of 1 being less than or equal to N being less than or equal to N at the moment of N
Figure GDA0003213770280000056
Wherein
Figure GDA0003213770280000057
The trace point sets are all under batch processing at the time of N' and come from the nth frame of the same target; and is initialized thereby
Figure GDA0003213770280000058
And Pnn(1≤n≤N)。
Wherein the content of the first and second substances,
Figure GDA0003213770280000061
indicating the updated state estimate, P, at the nth framennRepresents the updated covariance estimate at the nth frame, and t represents the trace of points
Figure GDA0003213770280000062
The sequence number of the echo data of the frame is in the echo data of the frame number L of the whole total observation data plane;
the processing procedure of steps 4-5 is shown in fig. 2:
step 4, updating the trace point set
At the moment of k +1, the echo data of the (k-N + 2) ·, k +1 frame is read from the radar receiver, and a track point sequence is output by a tracking algorithm before detection
Figure GDA0003213770280000063
Thereby obtaining an updated trace point set
Figure GDA0003213770280000064
k-N +2 is not less than N is not less than k +1, and
Figure GDA0003213770280000065
step 5, for N ═ k-N +2
Figure GDA0003213770280000066
And (3) carrying out filtering treatment:
step 5.1, firstly, a time point set with N being equal to k-N +2
Figure GDA0003213770280000067
Inner part
Figure GDA0003213770280000068
Spot trace
Figure GDA0003213770280000069
Performing Kalman filtering to obtain corresponding updated state estimation
Figure GDA00032137702800000610
Sum covariance estimation
Figure GDA00032137702800000611
Step 5.1.1 Aligning
Figure GDA00032137702800000612
Performing Kalman filtering to obtain corresponding updated state estimation
Figure GDA00032137702800000613
Sum covariance estimation
Figure GDA00032137702800000614
Including the following calculations:
intermediate variables
Figure GDA00032137702800000615
Intermediate variables
Figure GDA00032137702800000616
Wherein (·)TFor matrix transposition
Intermediate variables
Figure GDA00032137702800000617
Intermediate variables
Figure GDA00032137702800000618
Updated covariance estimation
Figure GDA00032137702800000619
Step 5.1.2, judging whether the requirement is met
Figure GDA00032137702800000620
If so, then there is an updated state estimate
Figure GDA00032137702800000621
Otherwise, there is
Figure GDA00032137702800000622
Step 5.1.3, judging whether the requirement is met
Figure GDA00032137702800000623
If yes, go to step 5.2; otherwise let t ═ t +1 and go to step 5.1.1.
Step 5.2, estimating all update states at the moment when N is k-N +2 according to determinant minimum covariance criterion
Figure GDA00032137702800000624
Sum covariance estimation
Figure GDA00032137702800000625
Solving the optimal combination:
Figure GDA0003213770280000071
Figure GDA0003213770280000072
obtaining corresponding weight vector
Figure GDA0003213770280000073
And solving to obtain the fused updated state estimation
Figure GDA0003213770280000074
Sum covariance estimation Pnn：
Figure GDA0003213770280000075
Figure GDA0003213770280000076
Wherein the content of the first and second substances,
Figure GDA0003213770280000077
representing covariance estimates
Figure GDA0003213770280000078
The weight of (2);
step 5.3, judging whether n is equal to k +1 or not, if yes, turning to step 6; otherwise, let n be n +1 and go to step 5.1.
Step 6, judging whether k +1 is equal to L or not, and if so, finishing updating all target states in the total observation time L frame; otherwise, k is made to be k +1, and the step is switched to.
FIG. 3 is a graph comparing the effect of the solution of the present invention and the conventional method; perfect Filtering in FIG. 3 refers to directly Filtering the real position of the target; the conventional DP-TBD is a Traditional DP-TBD algorithm (namely, if the detected track segment sequence is within a certain error range with the target track, the last frame point track of the track segment sequence is used for updating the target track); KF-CI-DP-TBD is the method provided by the invention, and Traditional DP-TBD is the Traditional processing method, and as can be seen from FIG. 3(a), the estimated track of the KF-CI-DP-TBD is closer to the real track than the Traditional DP-TBD in a single scene; as can be seen from FIG. 3(b), the RMSE (Root Mean Square Error) performance of the KF-CI-DP-TBD of the invention is better than that of the conventional DP-TBD; as can be seen by combining FIG. 3(b) and FIG. 3(c), as the SNR of the target increases, the RMSE of the KF-CI-DP-TBD gradually approaches the RMSE obtained by directly filtering the real position of the target, which shows the robustness of the KF-CI-DP-TBD algorithm, and shows that the method of the present invention has high and stable estimation accuracy.
It will be appreciated by those of ordinary skill in the art that the embodiments described herein are intended to assist the reader in understanding the principles of the invention and are to be construed as being without limitation to such specifically recited embodiments and examples. Various modifications and alterations to this invention will become apparent to those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the claims of the present invention.

Claims (4)

1. A real-time filtering method for output trace sequences of a multi-frame joint detection and tracking algorithm is characterized by comprising the following steps:
s1, obtaining a trace point set belonging to the same target at the same time according to a trace point sequence output by a tracking algorithm before batch processing detection;
s2, respectively performing Kalman filtering on the point traces in the point trace set of the same target at the same time obtained in the step S1, and obtaining the updated state estimation and the posterior covariance estimation of all the point traces in the point trace set of the same target at the same time;
s3, solving the minimum value of the updated state estimation and the posterior covariance estimation of all the point traces in the point trace set at the same time of the same target based on the determinant minimization criterion to obtain the weight corresponding to the updated posterior covariance estimation of all the point traces in the point trace set at the same time of the same target;
and S4, obtaining the final state estimation and the posterior covariance estimation through weighted summation according to the weight values corresponding to all the updated posterior covariance estimations obtained in the step S3 at each moment.
2. The method as claimed in claim 1, wherein the step S2 comprises the following steps:
if it is
Figure FDA0003213770270000011
Then adopt
Figure FDA0003213770270000012
Updating the state estimation of the trace point; otherwise adopt
Figure FDA0003213770270000013
Updating the state estimation of the trace point;
wherein, (.)TPerforming matrix transposition operation; (.)-1Performing matrix inversion operation; thetan tRepresenting a trace of points; t represents a trace of dots
Figure FDA0003213770270000014
The sequence number of the echo data of the frame is in the echo data of the frame number of the total observation data plane; h represents an observation matrix; gamma represents the elliptic wave gate threshold; f represents a target state transition matrix; q represents process noise covariance;
Figure FDA0003213770270000015
Figure FDA0003213770270000016
is an intermediate variable;
the above-mentioned
Figure FDA0003213770270000017
Is updated by the expression of
Figure FDA0003213770270000018
Figure FDA0003213770270000019
Represents the updated state estimate at frame n-1;
the above-mentioned
Figure FDA00032137702700000110
Is updated by the expression of
Figure FDA00032137702700000111
Figure FDA00032137702700000112
Is an intermediate variable, and the update expression is:
Figure FDA00032137702700000113
Rt-n+1represents the point trace measurement thetan tA corresponding measured noise covariance; pn-1n-1Represents the updated covariance estimate at frame n-1;
the updating covariance estimation in step S2 specifically includes:
Figure FDA00032137702700000114
Figure FDA00032137702700000115
representing an intermediate variable;
Figure FDA00032137702700000116
representing an updated covariance estimate.
3. The method as claimed in claim 1, wherein the step S1 is specifically as follows: and at the current moment, reading continuous N frames of echo data cut off to the current moment from the radar receiver, and outputting a trace point sequence by a track-before-detect algorithm so as to obtain a trace point set of the same target at the same moment.
4. The method as claimed in claim 3, wherein the step S1 is preceded by initializing the state estimate and the a posteriori covariance estimate, specifically: at the Nth' moment, the 1 st, 2 nd, 9th, N frames of echo data are read from the radar receiver, and a track point sequence is output by a track-before-detection algorithm
Figure FDA0003213770270000021
Thus obtaining the point trace set from all frames with the frame number of 1 being less than or equal to N being less than or equal to N at the N' th moment
Figure FDA0003213770270000022
Initializing state estimation and posterior covariance estimation according to the state estimation and the posterior covariance estimation;
wherein the content of the first and second substances,
Figure FDA0003213770270000023
set of traces, θ, for all batches at time N' and from the nth frame of the same targetn tRepresents the trace measurement, t represents the trace
Figure FDA0003213770270000024
The sequence number of the echo data of the frame in the echo data of the frame number L of the whole total observation data plane.
CN201910961671.7A 2019-10-11 2019-10-11 Real-time filtering method for trace point sequence for multi-frame joint detection and tracking Active CN110673132B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910961671.7A CN110673132B (en) 2019-10-11 2019-10-11 Real-time filtering method for trace point sequence for multi-frame joint detection and tracking

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910961671.7A CN110673132B (en) 2019-10-11 2019-10-11 Real-time filtering method for trace point sequence for multi-frame joint detection and tracking

Publications (2)

Publication Number Publication Date
CN110673132A CN110673132A (en) 2020-01-10
CN110673132B true CN110673132B (en) 2022-01-11

Family

ID=69081402

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910961671.7A Active CN110673132B (en) 2019-10-11 2019-10-11 Real-time filtering method for trace point sequence for multi-frame joint detection and tracking

Country Status (1)

Country Link
CN (1) CN110673132B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111413693B (en) * 2020-04-10 2022-02-11 中国人民解放军海军航空大学 TBD (tunnel boring device) and conventional tracking combination method based on double-threshold shunt processing in MIMO (multiple input multiple output) radar

Family Cites Families (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
FR2939904B1 (en) * 2008-12-16 2015-08-07 Thales Sa METHOD OF PROCESSING MULTI-SENSOR TRACKING WITH REDUCED LATENCY TIME
JP2011185719A (en) * 2010-03-08 2011-09-22 Mitsubishi Electric Corp Tracking radar device
CN102608568B (en) * 2012-02-28 2014-01-29 杭州电子科技大学 OSPA (Optimal Subpattern Assignment) distance track correlation method with fixed sliding window
CN103885057B (en) * 2014-03-20 2016-08-24 西安电子科技大学 Adaptive strain sliding window multi-object tracking method
CN104237853B (en) * 2014-09-16 2016-09-28 电子科技大学 A kind of for the particle filter method of trace point mark sequence before multi frame detection
CN105487061B (en) * 2015-12-01 2018-04-17 中国人民解放军海军航空工程学院 The multicharacteristic information fusion method of target data association
CN106226750B (en) * 2016-07-01 2018-06-19 电子科技大学 A kind of point mark sequence smooth filtering method for multi-frame joint detection
CN107656265B (en) * 2017-09-19 2021-03-30 电子科技大学 Particle filter fusion method for tracking short flight path before multi-frame detection
CN108333571B (en) * 2018-02-07 2020-04-21 电子科技大学 Multi-sensor multi-frame joint detection tracking method based on trace point sequence fusion
CN109633628B (en) * 2019-01-07 2023-04-21 西安电子科技大学 RGPO interference resisting method based on distributed networking radar data fusion

Also Published As

Publication number Publication date
CN110673132A (en) 2020-01-10

Similar Documents

Publication Publication Date Title
CN109946671B (en) Underwater maneuvering weak target detection and tracking method based on double-threshold judgment
CN109031277B (en) Multi-target image domain steady tracking method for through-wall radar
CN111580084B (en) Multi-distance extended target-oriented multi-Bernoulli pre-detection tracking method
Clark et al. GM-PHD filter multitarget tracking in sonar images
CN107356922B (en) Multi-frequency radar target tracking method for ambiguity resolution by spread-table-aided method
CN107346020B (en) Distributed batch estimation fusion method for asynchronous multi-base radar system
CN111157953B (en) Two-stage threshold constant false alarm detection algorithm under strong ground clutter
CN110865343A (en) LMB-based particle filter tracking-before-detection method and system
CN110673132B (en) Real-time filtering method for trace point sequence for multi-frame joint detection and tracking
CN111178261A (en) Face detection acceleration method based on video coding technology
CN108010066B (en) Multi-hypothesis tracking method based on infrared target gray level cross-correlation and angle information
CN105974402B (en) Suitable for tracking before the maneuvering target multi frame detection of pulse Doppler radar
CN107329131B (en) Radar weak target detection tracking method using particle filtering
CN106033120B (en) A kind of asynchronous multi-frame joint detection method of multistation radar
JP4196684B2 (en) Target tracking device
CN108845299B (en) Multi-sensor multi-frame joint detection algorithm based on posterior information fusion
Zhang et al. Adaptive Gaussian mixture probability hypothesis density for tracking multiple targets
Park et al. Track detection of low observable targets using a motion model
CN113608190B (en) Sea surface target detection method and system based on three characteristics of singular space
Wang et al. A computationally efficient dynamic programming based track-before-detect
Meng et al. A dynamic programming track-before-detect algorithm based on EKF for acceleration targets
Macagnano et al. Gating for multitarget tracking with the gaussian mixture phd and cphd filters
CN111796267A (en) Maneuvering turning target tracking-before-detection method based on pseudo-spectrum matched filtering
CN113192110A (en) Multi-target tracking method, device, equipment and storage medium
Xin et al. Object tracking using meanshift algorithm combined with Kalman filter on robotic fish

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