CN105929390B - Tracking before a kind of multi-target detection - Google Patents
Tracking before a kind of multi-target detection Download PDFInfo
- Publication number
- CN105929390B CN105929390B CN201610241167.6A CN201610241167A CN105929390B CN 105929390 B CN105929390 B CN 105929390B CN 201610241167 A CN201610241167 A CN 201610241167A CN 105929390 B CN105929390 B CN 105929390B
- Authority
- CN
- China
- Prior art keywords
- state
- value function
- frame
- value
- tracking
- 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
Links
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
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
Tracking before a kind of multi-target detection of the disclosure of the invention, belongs to Radar Technology field, is related to target detection tracking method, available for the radar dim target detection tracking processing in very noisy/clutter environment.For value function scaling problem existing for existing dynamic programming algorithm, the present invention proposes track algorithm before a kind of new multi-target detection, the algorithm uses new value function calculation, eliminate value function expansion, reduce false track number so as to reach, improve purpose of the radar to the detecting and tracking performance of approaching target.It is modified by calculating value function, has effectively eliminated value function expansion, thereby reduced the value function number of thresholding, so as to reduce the bar number of false track, and improved the detecting and tracking performance of approaching target.
Description
Technical field
The invention belongs to Radar Technology field, it is related to target detection tracking method, available in very noisy/clutter environment
The tracking of radar dim target detection is handled.
Background technology
With the increasingly complication of radar detection target and working environment, radar is difficult to effective detection such as long-range attack
Property target, the target being hidden in the strong clutter in ground sea, Stealthy Target etc., the increase of early radar warning time, power range reduce, to state
Family's safety brings serious threat.
The detecting and tracking to weak target is realized using signal processing mode, means are flexible, and cost is relatively low, application prospect
It is wide.Track algorithm before detection can be generally utilized, the more frame data of Combined Treatment, target energy is accumulated, improves echo
Signal to noise ratio, so as to improve detecting and tracking ability of the radar to weak target.
Track algorithm mainly includes before detection：Track algorithm before track algorithm, particle filter detection before Hough transformation detection
Track algorithm before being detected with Dynamic Programming.Wherein track algorithm carries out measuring value to possible targetpath before Dynamic Programming detection
Accumulation, if accumulating value (being referred to as value function) exceedes given thresholding, it is targetpath to assert this flight path.
However, track algorithm existence value function expansion problem, i.e. dbjective state and approaching target shape before Dynamic Programming detection
Value function of the value function apparently higher than remaining noise states corresponding to the noise states of state.Value function extension adds false-alarm flight path
Number, reduce detecting and tracking performance of the radar to approaching target.
The content of the invention
For value function scaling problem existing for existing dynamic programming algorithm, the present invention proposes a kind of new multiple target inspection
Track algorithm before survey, the algorithm use new value function calculation, eliminate value function expansion, reduce falseness so as to reach
Flight path number, improve purpose of the radar to the detecting and tracking performance of approaching target.
The technical scheme is that tracking before a kind of multi-target detection, this method comprise the following steps：
Step 1：Obtain echo-signal；
Step 2：Obtain per state space X corresponding to frame echo-signalkAnd the corresponding resolution of each state is single in state space
The measuring value of member；
Step 3：Value function initializes：By the measuring value z of the 1st each resolution cell of frame1(x1) it is assigned to the resolution cell pair
The value function I for the state answered1(x1)；
Step 4：Value function iteration：
Step 4.1：To kth frame data, the frame max function of kth -1 τ is obtained；
Step 4.2：Obtain state corresponding to the max function
Step 4.3：Searching stateIt is transferred to effective transfering state set of kth frame
Step 4.4：Set of computationsMaximum measuring value corresponding to middle stateThen look for maximum measurement
State corresponding to valueAssertIt is transferred toAnd recording status transfer relationship；
Step 4.5：Update value functionThen make
Step 4.6：Using step 4.1~step 4.5 identical method, shape corresponding to the 2nd frame to each state of k-th frame is determined
The renewal of state transfer relationship and corresponding value function；
Step 5：For k-th frame, the state that value function is more than appointed threshold is found；
Step 6：The state obtained for step 5, its corresponding targetpath is recalled according to the record of state transfer relationship；
Step 7：Obtain flight path and testing result.
The present invention is modified by calculating value function, has been effectively eliminated value function expansion, has been thereby reduced
The value function number of thresholding is crossed, so as to reduce the bar number of false track, and improves the detecting and tracking performance of approaching target.
Brief description of the drawings
Fig. 1 is the FB(flow block) of the present invention.
Fig. 2 is the true flight path of target.
Fig. 3 is that track algorithm handles obtained two-dimentional value function schematic diagram before conventional dynamic planning detects.
Fig. 4 is that track algorithm handles obtained targetpath schematic diagram before conventional dynamic planning detects, wherein the reality with "+"
Track algorithm recovers obtained flight path before line expression Dynamic Programming detection, and the solid line of band " o " represents real goal flight path.
Fig. 5 is the two-dimentional value function schematic diagram that present invention processing obtains.
Fig. 6 is the targetpath schematic diagram that present invention processing obtains, wherein the solid line with "+" represents that the present invention recovers to obtain
Flight path, band " o " solid line represent real goal flight path.
Embodiment
The present invention can effectively eliminate value function expansion than tracking before conventional dynamic planning detection, reduce
False track number, improve the detecting and tracking performance to approaching target.
It is of the invention mainly to be verified that all steps, conclusion are all on MATLAB R2012b using the method for l-G simulation test
Checking is correct.Specific implementation step is as follows：
Step 1：Systematic parameter initializes.Range cell number Nr=30, localizer unit number Nθ=30, distance dimension is most
Small speed V1=-1, the maximal rate V of distance dimension2=1, the minimum speed V of azimuth dimension3=-2, the maximal rate V of azimuth dimension4=
2, accumulate frame number K=6, state transfer number q=9, detection threshold VT=14.1463.
Step 2：Establish the state set per frame.For kth (1≤k≤K) frame, state set is expressed as Xk=[i, r,
j,s]′,1≤i≤Nr,1≤j≤Nθ,V1≤r≤V2,V2≤s≤V2, wherein i represent range direction on position, r represent away from
From the speed on direction, j represents the position on azimuth direction, and s represents the speed on azimuth direction.
Step 3：Value function initializes.For each state x of the 1st frame1=[i, r, j, s] ' ∈ X1, to being worth corresponding to it
Function I1(x1) assignment, i.e. I1(x1)=z1(x1), wherein z1(x1) represent the 1st frame resolution cell (i, j) in measuring value.And make
S1(x1)=0, wherein S1(x1) recording status transfer relationship, i.e. state x1In state corresponding to preceding 1 frame.
Step 4：Value function iteration.
Step 4.1：For 2≤k≤K, the max function of the frame of kth -1 is found, i.e.,
Step 4.2：Obtain state corresponding to the max functionI.e.
Step 4.3：Searching stateIt is transferred to effective transfering state set of kth frame
Step 4.4：Set of computationsMaximum measuring value corresponding to middle state, i.e.,Then seek
Look for state corresponding to the maximum measuring valueI.e.AssertIt is transferred toAnd recording status turns
Shifting relation, i.e.,
Step 4.5：Update value functionThen make
Step 4.6：Using step 4.1~step 4.5 identical method, determine corresponding to the 2nd frame to each state of k-th frame
The renewal of state transfer relationship and corresponding value function；
Step 5：Detection judgement.For k-th frame, find value function and be more than appointed threshold VTState, i.e.,
Step 6：Flight path is recalled.The state obtained for step 5, according to corresponding to the record of state transfer relationship recalls it
Targetpath, i.e.,：For k=K-1, K-2 ..., 1,The flight path for then recalling to obtain is
Step 7：Flight path and testing result output.
As shown in figure 3, the value function of dbjective state before conventional dynamic planning detection after track algorithm processing and neighbouring
The value function of the noise states of dbjective state is significantly greater than the value function of remaining noise states, and here it is value function expansion.
As shown in figure 4, many false tracks are produced after tracking processing before conventional dynamic planning detection, and approaching target is not easy
Correct tracking (in this emulation, target 2 is not tracked correctly).By Fig. 5, it can be seen that, the present invention can effectively disappear
Except value function expansion, more precisely after the present invention is handled, the value function of dbjective state is significantly greater than noise states
Value function, it is achieved thereby that highlight target, suppress noise, improve the purpose of echo signal to noise ratio.It is of the invention as seen from Figure 6
Can effectively reduce the number of false track and improve approaching target detecting and tracking performance (in this emulation, target 2
To correct tracking).
Claims (1)
1. tracking before a kind of multi-target detection, this method comprise the following steps：
Step 1：Obtain echo-signal；
Step 2：Obtain per state space X corresponding to frame echo-signalkAnd each state corresponds to the amount of resolution cell in state space
Measured value；
Step 3：Value function initializes；For each state x of the 1st frame1=[i, r, j, s] ' ∈ X1, to its corresponding value function
I1(x1) assignment, i.e. I1(x1)=z1(x1), wherein z1(x1) represent the 1st frame resolution cell (i, j) in measuring value；And make S1
(x1)=0, wherein S1(x1) recording status transfer relationship, i.e. state x1In state corresponding to preceding 1 frame, wherein i represents range direction
On position, r represent range direction on speed, j represent azimuth direction on position, s represent azimuth direction on speed；
Step 4：Value function iteration：
Step 4.1：To kth frame data, the frame max function τ of kth -1 is obtained,
Step 4.2：Obtain state corresponding to the max function
Step 4.3：Searching stateIt is transferred to effective transfering state set of kth frameWherein q represents state
Shift number；
Step 4.4：Set of computationsMaximum measuring value corresponding to middle state, i.e.,Then look for this
State corresponding to maximum measuring valueI.e.
AssertIt is transferred toAnd recording status transfer relationship, i.e.,
Step 4.5：Update value functionThen make
Step 4.6：Using step 4.1~step 4.5 identical method, determine that state corresponding to the 2nd frame to each state of k-th frame turns
The renewal of shifting relation and corresponding value function；
Step 5：For k-th frame, the state that value function is more than appointed threshold is found, i.e.,
Step 6：The state obtained for step 5, its corresponding targetpath is recalled according to the record of state transfer relationship, i.e.,：
For k=K-1, K-2 ..., 1,The flight path for then recalling to obtain isWherein Sk+1Represent,It is transferred toAnd recording status transfer relationship；
Step 7：Obtain flight path and testing result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610241167.6A CN105929390B (en) | 2016-04-18 | 2016-04-18 | Tracking before a kind of multi-target detection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610241167.6A CN105929390B (en) | 2016-04-18 | 2016-04-18 | Tracking before a kind of multi-target detection |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105929390A CN105929390A (en) | 2016-09-07 |
| CN105929390B true CN105929390B (en) | 2018-02-13 |
Family
ID=56838441
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610241167.6A Active CN105929390B (en) | 2016-04-18 | 2016-04-18 | Tracking before a kind of multi-target detection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105929390B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107340517B (en) * | 2017-07-04 | 2021-02-05 | 电子科技大学 | Multi-sensor multi-frame tracking-before-detection method |
| CN113721223A (en) * | 2021-08-04 | 2021-11-30 | 南京莱斯电子设备有限公司 | Tracking technology performance analysis method before detection of weak and small targets based on extreme value theory |
-
2016
- 2016-04-18 CN CN201610241167.6A patent/CN105929390B/en active Active
Non-Patent Citations (4)
| Title |
|---|
| "An Efficient Multi-Frame Track-Before-Detect Algorithm for Multi-Target Tracking";Wei Yi et al.;《IEEE JOURNAL OF SELECTED TOPICS IN SIGNAL PROCESSING》;20130630;第7卷(第3期);第421-434页 * |
| "Thresholding Process Based Dynamic Programming Track-Before-Detect Algorithm";Wei YI et al.;《IEICE TRANS.COMMUN.》;20130131;第E96-B卷(第1期);第291-300页 * |
| "基于动态规划的多目标GMTI-TBD技术研究";李渝 等;《仪器仪表学报》;20160229;第37卷(第2期);第356-364页 * |
| "机动弱小目标动态规划检测前跟踪方法";万洋 等;《信号处理》;20130531;第29卷(第5期);第584-590页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105929390A (en) | 2016-09-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103885057B (en) | Adaptive strain sliding window multi-object tracking method | |
| CN103439697B (en) | Target detection method based on dynamic programming | |
| CN103472445B (en) | Detecting tracking integrated method for multi-target scene | |
| CN103852756B (en) | Target detecting and tracking method by utilization of continuous wave radar | |
| CN104714225B (en) | Dynamic programming tracking-before-detection method based on generalized likelihood ratios | |
| CN106054169A (en) | Multi-station radar signal fusion detection method based on tracking information | |
| CN104297748A (en) | Before-radar-target-detecting tracking method based on track enhancing | |
| CN103729859A (en) | Probability nearest neighbor domain multi-target tracking method based on fuzzy clustering | |
| CN102243303A (en) | Method for locating static human body through wall based on respiratory character | |
| CN103885053A (en) | Radar data moving-target detection processing method based on tracking filter | |
| CN104237879A (en) | Multi-target tracking method in radar system | |
| CN105372659A (en) | Road traffic monitoring multi-target detection tracking method and tracking system | |
| CN104931934A (en) | Radar plot clotting method based on PAM clustering analysis | |
| CN106772352A (en) | A kind of PD radars extension Weak target detecting method based on Hough and particle filter | |
| CN103176168A (en) | Short-range cluster cancellation method for airborne non-side-looking array radar | |
| CN103777189A (en) | Radar weak target detecting method based on information geometry multiple autoregressive model | |
| CN103197297B (en) | Radar moving target detection method based on cognitive framework | |
| CN104898115A (en) | Multi-target tracking method after through-wall radar imaging | |
| CN108344982A (en) | Small drone target radar detection method based on long-time phase-coherent accumulation | |
| CN104155651A (en) | Probability data correlation method for polarizing radar target tracking | |
| CN103728599A (en) | Method for restraining deception false target jamming through remotely configured initiative and passive radar net | |
| CN105301584A (en) | IPPHDF maneuvering multi-target tracking method of simultaneously solving range ambiguities | |
| CN105044686A (en) | Radar dense false target interference inhibition method | |
| CN104865570A (en) | Rapid dynamic programming track-before-detect method | |
| CN107942324A (en) | Multi-frame joint Small object double check method based on Doppler's guiding |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |