CN110988833B - Weak target detection and tracking method - Google Patents
Weak target detection and tracking method Download PDFInfo
- Publication number
- CN110988833B CN110988833B CN201911148279.7A CN201911148279A CN110988833B CN 110988833 B CN110988833 B CN 110988833B CN 201911148279 A CN201911148279 A CN 201911148279A CN 110988833 B CN110988833 B CN 110988833B
- Authority
- CN
- China
- Prior art keywords
- target
- denotes
- state information
- moment
- value
- 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
- 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
- G01S7/418—Theoretical aspects
-
- 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
- G01S13/006—Theoretical aspects
-
- 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
-
- 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
- G01S7/414—Discriminating targets with respect to background clutter
-
- 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
- G01S7/415—Identification of targets based on measurements of movement associated with the target
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
The invention discloses a weak target detection and tracking method based on weighted dynamic programming and Kalman filtering, which utilizes multi-frame data obtained by a sensor to accumulate energy by using a dynamic programming algorithm and selects a path with the maximum energy as a predicted flight path, thereby solving the problem of huge calculated amount caused by adopting an exhaustion method. And when energy accumulation is carried out, weighting processing is carried out by adopting a weighting function, and the zero position of the weighting function is determined by combining with a predicted value given by Kalman filtering. The dynamic programming algorithm after the weighting function and the Kalman filtering processing can effectively solve the energy diffusion problem in the traditional dynamic programming algorithm, so that the accuracy of establishing the flight path is improved.
Description
Technical Field
The invention relates to a weak target detection and tracking method, and belongs to the technical field of target detection and target tracking.
Background
The tracking algorithm before detection is an algorithm which is mainly provided for solving the problem of detection and tracking of a weak target under a clutter background. The algorithm utilizes the sensor to scan for multiple times to obtain data containing noise and targets, joint processing is carried out on the data, and multiple groups of data are fully utilized, so that the weak targets are detected and tracked. In order to implement the pre-detection tracking algorithm, various methods have been proposed, such as a pre-detection tracking algorithm based on particle filtering, a pre-detection tracking algorithm based on dynamic programming, a pre-detection tracking algorithm based on hough transform, and the like.
The pre-detection tracking algorithm based on dynamic programming is an optimization algorithm of an exhaustive method, lists all possible paths, accumulates energy along the paths, and considers that the value obtained by accumulating the energy along the real track is the largest, so that the path with the larger energy accumulation value is more likely to be the real track. The dynamic programming algorithm can effectively reduce the calculated amount and avoid the defect that the calculated amount of an exhaustion method increases exponentially.
Although the pre-detection tracking algorithm based on dynamic programming can greatly reduce the calculation amount compared with an exhaustive method, the traditional dynamic programming algorithm does not limit the range of energy accumulation and does not consider the weight, so that the method has a more serious energy diffusion problem, and further more false tracks are generated.
Disclosure of Invention
The technical problem to be solved by the invention is to overcome the defects of the prior art and provide a weak target detection and tracking method.
In order to solve the technical problem, the invention provides a weak target detection and tracking method, which comprises the steps of obtaining the measured value of each resolution unit in an echo signal at the moment k, and obtaining target state information at the echo moment according to all the measured values at the moment;
initializing a value function and target state information at the moment k-1, wherein the value function represents the accumulated value of the echo signal;
updating the value function and the target state information when k is 2 according to the initialized value function and the target state information;
updating the target state information when k is 3 according to the value function when k is 2, and calculating the motion speed of the target when k is 2 by using the target state information when k is 3 and the target state information when k is 2;
when k is 3, updating the target state information of the current moment through the value function of the previous moment, obtaining the position information of the previous moment through the target state information of the previous moment, processing the motion speed and the position information of the target obtained at the previous moment by adopting a Kalman filtering and direction weighting method, and updating the value function of the current moment;
when k is larger than or equal to 4, a recursion formula is executed, and until the last frame data is processed, the recursion formula sequentially comprises the following steps: updating the record of the target state information at the current moment, estimating the target motion speed at the previous moment, performing Kalman filtering and direction weighting according to the target state information and the target motion speed at the previous moment, updating the search range of the target state information at the next moment, and updating the value function at the current moment;
intercepting the threshold value of the value function to obtain all the values meeting the preset threshold value condition
To the aboveAnd (4) tracing back the flight path by using the record of the target state information to obtain the flight path obtained at the kth moment.
Further, the measurement value is expressed as:
wherein s isij(k) Representing the echo signal of the target, nij(k) Representing a random noise signal, said noise signal obeying a standard normal distribution and, for different instants k, a random noise signal nij(k) Independently distributed, detected target state information is represented as xk(x, y) which contains the position of the object, xk(x, y) is for zij(k) The position information obtained after signal processing is denoted as xk=F(zij(k) I and j respectively represent the position coordinates of the resolution cell
Further, the value function and the target state information at the time when the initialization k is 1 are respectively expressed as:
Further, when k is 2, the value function is updated: q (x)2)=max[q(x1)]+zij(2) Then, the target state information is updated: r (2) ═ argmax [ q (x) ]1)]。
Further, when k is 3, the target state information is updated first:
R(3)=argmax[q(x2)]
and estimating the moving speed of the target when k is 2 by combining the target state information at the previous moment:
wherein, vx2Denotes the speed in the x direction, vy, when k is 22Denotes a speed in the y direction when k is 2, x (-) denotes an x coordinate of the extraction · y (-) denotes a y coordinate of the extraction · TTime of flightRepresenting two framesThe time interval in between.
Further, when k is 3, a kalman filter algorithm is executed to predict the target position when k is 3:
P(2)=E(4)
wherein X2 denotes an initial state quantity, P (2) an initial covariance matrix, E (4) denotes a 4-order identity matrix, A denotes a state transition matrix, B denotes an external control matrix, T denotes a matrix transpose, J denotes an external control quantity,denotes a covariance matrix prediction value when K is 3, Q denotes a process noise covariance matrix, S denotes innovation, H denotes an observation matrix, R denotes an observation noise covariance matrix, K denotes a gain,represents the estimated motion velocity and position information of the target at time k-3, where x' (x)3)、y'(x3)、vx'3、vy'3Is x (x)3)、 y(x3)、vx3、vy3The predicted value of (2);
updating the zero position of the directional weighting function, wherein the directional weighting function is:
search range D of update status information:
D={x|vminT<|x-x2|<vmaxt }, said vmaxAnd vminMaximum velocity and maximum velocity of the object movement
Small velocity, x representing all the positions of the targets to be detected in the search area D, x2Represents the determined target position when k is 2;
update k is a function of the value at time 3:
q(x3)=max[w(θ)q(x2)]+zij(3)
wherein F (z)ij(3))∈D。
Further, when k is larger than or equal to 4, the following steps are executed in a circulating mode until the processing of the last frame data is finished:
(1) and updating the record of the state information at the moment k:
R(k)=argmax[q(xk-1)]
wherein q (x)k-1) A value function representing a last time instant;
(2) estimating the speed:
(3) performing Kalman filtering on the record of the target state information at the current moment:
wherein X (k-1) represents a state quantity at the time of k-1,represents the predicted value of the target state information at the time k-1, P (k-1) represents the covariance matrix at the time k-1,represents a covariance matrix predicted value at the time k-1, E (4) represents an identity matrix of order 4, A represents a state transition matrix, B represents an external control matrix, T represents a matrix transposition, J represents an external control quantity,denotes the covariance matrix prediction value at time K, Q denotes the process noise covariance matrix, S denotes innovation, H denotes the observation matrix, R denotes the observation noise covariance matrix, K denotes gain,indicating the predicted value of the state at time k,represents the observed value at time k-1;
(4) directional weighting:
where θ is the predicted azimuth and (x)k-1),y(xk-1))、(x'(xk),y'(xk) Angle of the line connecting the two points;
(5) updating the search range of the next state information:
D={x|vminT<|x-xk-1|<vmaxT}
(6) update value function:
q(xk)=max[w(θ)q(xk-1)]+zij(k)
wherein z isij(k)∈D。
Further, threshold interception is carried out to obtain all the states of the Mth moment meeting the following conditions
Wherein V is a threshold value set according to a certain condition, and M is the total frame number.
Further, satisfy all of Mth frameOf the conditionAnd performing track backtracking by using the record of the state information, wherein the track obtained at the kth moment is recorded as TkThe method comprises the following steps:
the invention achieves the following beneficial effects:
the invention solves the problem of huge calculation amount caused by adopting an exhaustion method; when energy accumulation is carried out, weighting processing is carried out by adopting a weighting function, and the zero position of the weighting function is determined by combining a predicted value given by Kalman filtering; the dynamic programming algorithm after the weighting function and the Kalman filtering processing can effectively solve the energy diffusion problem in the traditional dynamic programming algorithm, thereby improving the accuracy of track establishment.
Detailed Description
The invention is further described below. The following examples are only for illustrating the technical solutions of the present invention more clearly, and the protection scope of the present invention is not limited thereby.
A weak target detection and tracking method comprises the following steps:
(1) the echo signal contains N2A resolution unit for recording the measured value of each resolution unit as zij(k) The measured values of all resolution cells at time k form a matrix, which is recorded as:
Z(k)={zij(k)}
(1)
wherein i is more than or equal to 0, j is more than or equal to N, zij(k) To resolve the measurement in cell (i, j), it can be expressed as:
wherein s isij(k) Representing the echo signal of the target, nij(k) Representing a random noise signal. It is assumed here that the noise signal follows a standard normal distribution and that for different time instants k, a random noise signal nij(k) Independently and equally distributed. Let the detected target state information be xk(x, y) which contains the position of the object, xk(x, y) is for zij(k) The position information obtained after signal processing is denoted as xk=F(zij(k))。
(2) Determining the maximum velocity v of the movement of an objectmaxAnd minimum velocity vminAnd constructing a value function and a state when k is 1Recording of information, and initializing all of it:
(3) when k is 2, updating the value function of each state information to be searched:
q(x2)=max[q(x1)]+zij(2)
(4)
the record of status information is then updated:
(4) when k is 3, firstly updating the record of the state information:
and combining the moving speed of the target when the last state position estimation k is 2:
(5) when k is 3, from the obtained x2And vx2,vy2Prediction of x using Kalman Filter Algorithm3And corresponding vx3,vy3. Firstly, initializing various parameters of a Kalman filtering algorithm: set its state transition matrixExternal control matrixObservation matrixObserving a noise covariance matrixProcess noise covariance matrixCovariance matrixInitial state quantityExternal control quantityThen a filtering algorithm is executed:
X(2)=X2,P(2)=E(4) (8)
where x' (x)3)、y'(x3)、vx'3、vy'3Is x (x)3)、y(x3)、vx3、vy3The predicted value of (2).
Using (x' (x)3),y'(x3) And (x)3),y(x3) Two points) update the zero point of the directional weighting function, w (θ) being:
because the time interval T between two frames is extremely short, the target can only move forwards and can not return within the time T, and therefore the value of w (theta) is zero when the angle is more than 90 degrees and less than or equal to 180 degrees. The kalman filter gives a predicted value (x' (x) when k is 33),y'(x3) Therefore, the target is considered to be along the line segment (x' (x))3),y'(x3))-(x(x2),y(x2) The likelihood of directional motion is greatest, i.e., θ is defined as the predicted azimuth and (x' (x))3),y'(x3))、(x(x2),y(x2) W (θ) is the largest when θ is 0.
Updating a search range D of state information, D ═ x | vminT<|x-x2|<vmaxT},
Finally, the update value function:
q(x3)=max[w(θ)q(x2)]+zij(3) (14)
wherein F (z)ij(3))∈D
(6) When k is more than or equal to 4, the following steps are executed in a circulating way until the processing of the last frame data is finished. Write the step as recursive:
(1) recording of update status information:
(2) estimating the speed:
(3) kalman filtering:
wherein X (k-1) represents a state quantity at the time of k-1,represents the predicted value of the target state information at the time k-1, P (k-1) represents the covariance matrix at the time k-1,represents a covariance matrix predicted value at the time k-1, E (4) represents an identity matrix of order 4, A represents a state transition matrix, B represents an external control matrix, T represents a matrix transposition, J represents an external control quantity,representing coordination of time kThe variance matrix predictor, Q represents the process noise covariance matrix, S represents innovation, H represents the observation matrix, R represents the observation noise covariance matrix, K represents gain,indicating the predicted value of the state at time k,represents the observed value at time k-1;
(4) directional weighting:
where θ is the predicted azimuth and (x)k-1),y(xk-1))、(x'(xk),y'(xk) Angle between the two points.
(5) Updating the search range of the next state information:
D={x|vminT<|x-xk-1|<vmaxT} (24)
(6) update value function:
q(xk)=max[w(θ)q(xk-1)]+zij(k) (25)
wherein z isij(k)∈D。
(7) Intercepting a threshold value to find out all the states of the Mth moment meeting the following conditions
Wherein V is a threshold value set according to a certain condition, and M is the total frame number.
(8) Satisfy all of the Mth frameOf the conditionAnd performing track backtracking by using the record of the state information, wherein the track obtained at the kth moment is recorded as TkThe method comprises the following steps:
the above description is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, several modifications and variations can be made without departing from the technical principle of the present invention, and these modifications and variations should also be regarded as the protection scope of the present invention.
Claims (4)
1. A weak target detection and tracking method is characterized in that,
acquiring the measured value of each resolution unit in the echo signal at the moment k, and acquiring target state information at the echo moment according to all the measured values at the moment;
initializing a value function and target state information at the moment k-1, wherein the value function represents the accumulated value of the echo signal;
updating the value function and the target state information when k is 2 according to the initialized value function and the target state information;
updating the target state information when k is 3 according to the value function when k is 2, and calculating the motion speed of the target when k is 2 by using the target state information when k is 3 and the target state information when k is 2;
when k is 3, updating the target state information of the current moment through the value function of the previous moment, obtaining the position information of the previous moment through the target state information of the previous moment, processing the motion speed and the position information of the target obtained at the previous moment by adopting a Kalman filtering and direction weighting method, and updating the value function of the current moment;
when k is larger than or equal to 4, a recursion formula is executed, and until the last frame data is processed, the recursion formula sequentially comprises the following steps: updating the record of the target state information at the current moment, estimating the target motion speed at the previous moment, performing Kalman filtering and direction weighting according to the target state information and the target motion speed at the previous moment, updating the search range of the target state information at the next moment, and updating the value function at the current moment;
intercepting the threshold value of the value function to obtain all the values meeting the preset threshold value condition
To the aboveTracing back the flight path by using the record of the target state information to obtain the flight path obtained at the kth moment;
the measurement values are expressed as:
wherein s isij(k) Representing the echo signal of the target, nij(k) Representing a random noise signal, said noise signal obeying a standard normal distribution and, for different instants k, a random noise signal nij(k) Independently distributed, detected target state information is represented as xk(x, y) which contains the position of the object, xk(x, y) is for zij(k) The position information obtained after signal processing is denoted as xk=F(zij(k) I and j respectively represent the position coordinates of the resolution unit;
the value function and the target state information at the moment when the initialization k is 1 are respectively expressed as:
when k is 2, update value function: q (x)2)=max[q(x1)]+zij(2) Then, the target state information is updated: r (2) ═ argmax [ q (x) ]1)];
When k is 3, the target state information is updated firstly:
R(3)=argmax[q(x2)]
and estimating the moving speed of the target when k is 2 by combining the target state information at the previous moment:
wherein, vx2Denotes the speed in the x direction, vy, when k is 22Denotes a speed in the y direction when k is 2, x (-) denotes an x coordinate of the extraction · y (-) denotes a y coordinate of the extraction · TTime of flightRepresenting the time interval between two frames;
when k is 3, executing a Kalman filtering algorithm, and predicting the target position when k is 3:
P(2)=E(4)
wherein X2 denotes an initial state quantity, P (2) an initial covariance matrix, E (4) denotes a 4-order identity matrix, A denotes a state transition matrix, B denotes an external control matrix, T denotes a matrix transpose, J denotes an external control quantity,denotes a covariance matrix prediction value when K is 3, Q denotes a process noise covariance matrix, S denotes innovation, H denotes an observation matrix, R denotes an observation noise covariance matrix, K denotes a gain,represents the estimated motion velocity and position information of the target at time k-3, where x' (x)3)、y'(x3)、vx'3、vy'3Is x (x)3)、y(x3)、vx3、vy3The predicted value of (2);
updating the zero position of the directional weighting function, wherein the directional weighting function is:
where θ is the predicted azimuth and (x)k-1),y(xk-1))、(x'(xk),y'(xk) Angle of the line connecting the two points;
search range D of update status information: d ═ x | vminT<|x-x2|<vmaxT }, said vmaxAnd vminThe maximum speed and the minimum speed of the target motion are adopted, x represents all the positions of the target to be detected in the search area D, and x2Represents the determined target position when k is 2;
update k is a function of the value at time 3:
q(x3)=max[w(θ)q(x2)]+zij(3)
wherein F (z)ij(3))∈D。
2. The weak target detecting and tracking method according to claim 1, wherein when k ≧ 4, the following steps are performed in a loop until the end frame data is processed:
(1) and updating the record of the state information at the moment k:
R(k)=argmax[q(xk-1)]
wherein q (x)k-1) A value function representing a last time instant;
(2) estimating the speed:
(3) performing Kalman filtering on the record of the target state information at the current moment:
wherein X (k-1) represents a state quantity at the time of k-1,represents the predicted value of the target state information at the time k-1, P (k-1) represents the covariance matrix at the time k-1,represents a covariance matrix predicted value at the time k-1, E (4) represents an identity matrix of order 4, A represents a state transition matrix, B represents an external control matrix, T represents a matrix transposition, J represents an external control quantity,denotes the covariance matrix prediction value at time K, Q denotes the process noise covariance matrix, S denotes innovation, H denotes the observation matrix, R denotes the observation noise covariance matrix, K denotes gain,indicating the predicted value of the state at time k,represents the observed value at time k-1;
(4) directional weighting:
where θ is the predicted azimuth and (x)k-1),y(xk-1))、(x'(xk),y'(xk) Angle between the two points;
(5) updating the search range of the next state information:
D={x|vminT<|x-xk-1|<vmaxT}
(6) update value function:
q(xk)=max[w(θ)q(xk-1)]+zij(k)
wherein z isij(k)∈D。
4. A weak target detecting and tracking method according to claim 3, characterized in that all the fulfils in the mth frame are assignedOf the conditionAnd performing track backtracking by using the record of the state information, wherein the track obtained at the kth moment is recorded as TkThe method comprises the following steps:
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911148279.7A CN110988833B (en) | 2019-11-21 | 2019-11-21 | Weak target detection and tracking method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911148279.7A CN110988833B (en) | 2019-11-21 | 2019-11-21 | Weak target detection and tracking method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110988833A CN110988833A (en) | 2020-04-10 |
CN110988833B true CN110988833B (en) | 2022-05-20 |
Family
ID=70085685
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201911148279.7A Active CN110988833B (en) | 2019-11-21 | 2019-11-21 | Weak target detection and tracking method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110988833B (en) |
Family Cites Families (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP5398288B2 (en) * | 2009-02-05 | 2014-01-29 | 株式会社東芝 | Radar signal processing apparatus and target judgment method thereof |
CN104297748B (en) * | 2014-10-20 | 2017-03-08 | 西安电子科技大学 | One kind is based on tracking before the enhanced Radar Targets'Detection in track |
CN105223570B (en) * | 2015-10-22 | 2017-11-28 | 中国船舶重工集团公司第七二四研究所 | One kind is to track method for optimizing before extra large Dim targets detection |
CN106842164B (en) * | 2017-03-14 | 2019-06-14 | 中国人民解放军国防科学技术大学 | Non- cooperation pulse compression radar Weak target detecting method based on Wavelet Denoising Method |
CN108802707B (en) * | 2018-08-31 | 2021-03-26 | 中国科学院电子学研究所 | Improved Kalman filtering method for target tracking |
CN109946671B (en) * | 2019-04-12 | 2020-11-03 | 哈尔滨工程大学 | Underwater maneuvering weak target detection and tracking method based on double-threshold judgment |
-
2019
- 2019-11-21 CN CN201911148279.7A patent/CN110988833B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN110988833A (en) | 2020-04-10 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109946671B (en) | Underwater maneuvering weak target detection and tracking method based on double-threshold judgment | |
CN109100714B (en) | Low-slow small target tracking method based on polar coordinate system | |
CN111127518B (en) | Target tracking method and device based on unmanned aerial vehicle | |
CN110390292B (en) | Remote sensing video vehicle target detection and tracking method based on dynamic correlation model | |
WO2012009947A1 (en) | Device and method of signature-driven multi-target tracking | |
CN112991391A (en) | Vehicle detection and tracking method based on radar signal and vision fusion | |
CN112308881A (en) | Ship multi-target tracking method based on remote sensing image | |
CN110286357B (en) | Target motion positioning method based on underwater sound detection | |
CN111739053B (en) | Online multi-pedestrian detection tracking method under complex scene | |
CN104793200B (en) | Dynamic planning track-before-detect method based on iterative processing | |
CN108010066B (en) | Multi-hypothesis tracking method based on infrared target gray level cross-correlation and angle information | |
CN111754545A (en) | Dual-filter video multi-target tracking method based on IOU matching | |
CN109814074B (en) | Group target tracking method based on image morphological processing | |
CN115840221B (en) | Method for achieving target feature extraction and multi-target tracking based on 4D millimeter wave radar | |
CN116128932A (en) | Multi-target tracking method | |
CN110133612A (en) | A kind of extension target detection method based on tracking feedback | |
CN110717934A (en) | Anti-occlusion target tracking method based on STRCF | |
CN111259332B (en) | Fuzzy data association method and multi-target tracking method in clutter environment | |
CN106291530B (en) | A kind of probabilistic data association optimization method based on nearest neighbor method | |
CN117169818B (en) | Radar waveform design method for sea surface maneuvering target tracking | |
CN110988833B (en) | Weak target detection and tracking method | |
CN108828584B (en) | Multi-frequency target tracking-before-detection method based on track folding factor ambiguity resolution | |
CN115220002B (en) | Multi-target data association tracking method and related device for fixed single station | |
CN116381672A (en) | X-band multi-expansion target self-adaptive tracking method based on twin network radar | |
CN113126086B (en) | Life detection radar weak target detection method based on state prediction accumulation |
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 |