CN111142085B - External radiation source radar target classification and identification method based on track feature extraction - Google Patents

External radiation source radar target classification and identification method based on track feature extraction Download PDF

Info

Publication number
CN111142085B
CN111142085B CN202010040144.5A CN202010040144A CN111142085B CN 111142085 B CN111142085 B CN 111142085B CN 202010040144 A CN202010040144 A CN 202010040144A CN 111142085 B CN111142085 B CN 111142085B
Authority
CN
China
Prior art keywords
sub
target
track
classified
flight path
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
CN202010040144.5A
Other languages
Chinese (zh)
Other versions
CN111142085A (en
Inventor
万显荣
占伟杰
易建新
刘斯齐
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.)
Wuhan University WHU
Original Assignee
Wuhan University WHU
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 Wuhan University WHU filed Critical Wuhan University WHU
Priority to CN202010040144.5A priority Critical patent/CN111142085B/en
Publication of CN111142085A publication Critical patent/CN111142085A/en
Application granted granted Critical
Publication of CN111142085B publication Critical patent/CN111142085B/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
    • 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
    • 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
    • G01S7/418—Theoretical aspects

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Radar Systems Or Details Thereof (AREA)

Abstract

The invention discloses an external radiation source radar target classification and identification method based on track feature extraction, which constructs feature quantities including motion states, echo features, track expression forms and the like, and fully excavates the feature differences of different types of target tracks. The method comprises the steps of firstly extracting sub tracks from original tracks to be classified, calculating values of all characteristic quantities of the sub tracks, then analyzing various target tracks of known types collected in advance, obtaining typical ranges of all characteristic quantities corresponding to each type of targets, and finally calculating confidence degrees of the tracks to be classified corresponding to each type of targets according to the value conditions of all characteristic quantities of the tracks to be classified, so as to realize target classification and identification. Each step of the method has clear physical meaning, good interpretability, small calculated amount, easy engineering realization and popularization and application value.

Description

External radiation source radar target classification and identification method based on track feature extraction
Technical Field
The invention relates to the technical field of external radiation source radar signal processing, in particular to an external radiation source radar target classification and identification method based on track feature extraction.
Background
The external radiation source radar is a new system radar for realizing target detection by utilizing an electromagnetic signal transmitted by a third party, is concerned about due to the advantages of spectrum saving, environmental protection, easy networking and the like, and has great application potential in the fields of low-altitude monitoring, urban security and the like. The detection structure of the external radiation source radar transmitting and receiving split position provides more freedom degrees for collecting targets, and meanwhile, the detection sensitivity of the external radiation source radar is greatly improved by a long-time coherent accumulation technology. Therefore, the external radiation source radar can simultaneously detect high-altitude flight targets such as civil aircrafts and navigation aircrafts, low-speed and small targets such as unmanned planes and birds, and moving targets such as ground vehicles and pedestrians. Different types of targets may form similar motion tracks on a radar display terminal, so that judgment and extraction of key target information are influenced, and the application range of the external radiation source radar is limited. The classification identification of radar targets is beyond the basic radar functions of target detection, target tracking and the like, and has higher requirements on radars, thereby having great application value. The traditional method for distinguishing different types of targets by means of manual judgment of radar operators is time-consuming and labor-consuming, and is easy to have the situations of false alarm, missing report and the like, so that the development of an external radiation source radar target classification and identification algorithm is urgent.
The inventor of the present application finds that the method of the prior art has at least the following technical problems in the process of implementing the present invention:
(1) radar target classification method based on target micro-Doppler characteristics
The micro-doppler is a frequency modulation phenomenon caused by micro-motions such as rotation, vibration and the like except the motion of a target body, is a unique motion characteristic of the target, and can be introduced into the micro-doppler phenomenon by an engine of an airplane, a propeller of an unmanned aerial vehicle, wing vibration of a bird and the like. Although the classification and identification by using the micro-doppler phenomenon of the target is an effective method, the conventional external radiation source radar usually works in VHF and UHF frequency bands, and the micro-doppler phenomenon of the targets such as birds and unmanned planes is difficult to capture, so that the application of the method in the external radiation source radar is greatly restricted.
(2) Radar target classification method based on machine learning
According to the method, parameters such as motion characteristics, echo energy characteristics and track morphological characteristics of the target are extracted and used as input of machine learning, and classification and identification of the radar target are achieved by utilizing strong data fitting capacity of the machine learning. The machine learning algorithm needs to use a large number of training samples to train the model, the calculation amount is large, machine learning is just like a black box, the physical mechanism for realizing classification and identification is difficult to clearly understand, and when the number of samples is insufficient or excessive, the condition of under-fitting or over-fitting of the model is easy to occur, and the generalization capability of the model is difficult to ensure.
Therefore, the method in the prior art has the technical problems of large calculation amount and poor explanation.
Disclosure of Invention
In view of the above, the present invention provides a classification and identification method for an external radiation source radar target based on track feature extraction, so as to solve or at least partially solve the technical problems of large calculation amount and lack of interpretability existing in the methods in the prior art.
In order to solve the technical problem, the invention provides an external radiation source radar target classification and identification method based on track feature extraction, which comprises the following steps:
s1: acquiring an original flight path of a target to be classified, extracting a sub-flight path from the original flight path of the target to be classified, and calculating each characteristic quantity of the target to be classified according to the extracted sub-flight path;
s2: counting typical values of each characteristic quantity of each type of target from actually measured track data;
s3: determining a target typical characteristic quantity for rough classification according to typical values of various characteristic quantities of each type of target, wherein the difference between the typical value of the target typical characteristic quantity on the first type of target and the typical values of other types of targets on the target typical characteristic quantity is larger than a preset value, and the first type of target and the other types of targets are different types;
s4: roughly classifying the original flight paths of the targets to be classified according to the determined target typical characteristic quantities, screening out original flight paths of target types corresponding to the target typical characteristic quantities, and taking the original flight paths of the target types not corresponding to the target typical characteristic quantities as the residual target flight paths to be classified;
s5: and calculating the confidence degrees of the residual target tracks to be classified corresponding to various types of targets according to the relation between each characteristic value of the residual target tracks to be classified and the typical value of each characteristic quantity, and classifying the residual target tracks to be classified according to the confidence degrees.
In one embodiment, S1 specifically includes:
s1.1: acquiring an original track of a target to be classified;
s1.2: carrying out sliding windowing operation on the original track of the target to be classified, forming track points contained in each sliding window into sub tracks of the original track, and extracting the sub tracks;
s1.3: calculating each characteristic quantity of the target to be classified from the extracted sub-tracks, wherein the mth characteristic quantity is represented as Fm,m∈[1,M]Then F ism,m∈[1,M]Comprises the following steps:
(1)F1: average distance of sub-tracks
Figure BDA0002367446100000031
Wherein L issubRepresenting the length of each sub-track, NsubThe number of word tracks which can be extracted from the original track with the length L is represented,
Figure BDA0002367446100000032
represents the ith, i ∈ [1, N ]sub]Average distance of flight path of strip, rl (i)Represents the ith, i ∈ [1, N ]sub]The L, L in the flight path of the strip belongs to [1, L ∈ ]sub]The distance of each waypoint from the receiving station;
(2)F2: sub track average velocity
Figure BDA0002367446100000033
Wherein,
Figure BDA0002367446100000034
represents the ith, i ∈ [1, N ]sub]The average speed of the flight path of the strip,
Figure BDA0002367446100000035
represents the ith, i ∈ [1, N ]sub]The L, L in the flight path of the strip belongs to [1, L ∈ ]sub]The speed of each track point;
(3)F3: sub track average acceleration
Figure BDA0002367446100000036
Wherein,
Figure BDA0002367446100000037
represents the ith, i ∈ [1, N ]sub]The average speed of the flight path of the strip,
Figure BDA0002367446100000038
represents the ith, i ∈ [1, N ]sub]The L, L in the flight path of the strip belongs to [1, L ∈ ]sub]The speed of each track point;
(4)F4: average height of sub-track
Figure BDA0002367446100000039
Wherein,
Figure BDA00023674461000000310
represents the ith, i ∈ [1, N ]sub]The average height of the flight path of the strip,
Figure BDA00023674461000000311
represents the ith, i ∈ [1, N ]sub]The L, L in the flight path of the strip belongs to [1, L ∈ ]sub]The height of each course point;
(5)F5: sub-track target equivalent mean RCS: cross section of radar scattering
Figure BDA00023674461000000312
Wherein,
Figure BDA00023674461000000313
represents the ith, i ∈ [1, N ]sub]The average equivalent RCS of the strip flight path,
Figure BDA00023674461000000314
represents the ith, i ∈ [1, N ]sub]The L, L in the flight path of the strip belongs to [1, L ∈ ]sub]The equivalent RCS of each track point, A represents a constant related to system noise coefficient, signal transmitting power, receiving and transmitting antenna gain and other parameters, RrIndicating the distance, R, of the target from the receiving radar stationtRepresenting the distance of the target from the transmitting station, λ representing the wavelength of the transmitted signal, and SNR representing the target signal-to-noise ratio;
(6)F6: sub track velocity variance
Figure BDA0002367446100000041
Wherein,
Figure BDA0002367446100000042
represents the ith, i ∈ [1, N ]sub]The variance of the speed of the flight path of the strip,
Figure BDA0002367446100000043
represents the ith, i ∈ [1, N ]sub]The L, L in the flight path of the strip belongs to [1, L ∈ ]sub]The speed of each of the track points is,
Figure BDA0002367446100000044
represents the ith, i ∈ [1, N ]sub]Average speed of the flight path of the sliver;
(7)F7: sub track acceleration variance
Figure BDA0002367446100000045
Wherein,
Figure BDA0002367446100000046
represents the ith, i ∈ [1, N ]sub]The variance of the acceleration of the strip flight path,
Figure BDA0002367446100000047
represents the ith, i ∈ [1, N ]sub]The L, L in the flight path of the strip belongs to [1, L ∈ ]sub]The acceleration of each of the waypoints is,
Figure BDA0002367446100000048
represents the ith, i ∈ [1, N ]sub]Average acceleration of the strip flight path;
(8)F8: sub track height variance
Figure BDA0002367446100000049
Wherein,
Figure BDA00023674461000000410
represents the ith, i ∈ [1, N ]sub]The variance of the height of the flight path of the strip,
Figure BDA00023674461000000411
represents the ith, i ∈ [1, N ]sub]The L, L in the flight path of the strip belongs to [1, L ∈ ]sub]The height of each course point is determined,
Figure BDA00023674461000000412
represents the ith, i ∈ [1, N ]sub]Average height of the flight path;
(9)F9: sub-track equivalent RCS variance
Figure BDA00023674461000000413
Wherein,
Figure BDA00023674461000000414
represents the ith, i ∈ [1, N ]sub]The equivalent RCS of the strip flight path,
Figure BDA00023674461000000415
represents the ith, i ∈ [1, N ]sub]The L, L in the flight path of the strip belongs to [1, L ∈ ]sub]The equivalent RCS of each track point,
Figure BDA00023674461000000416
represents the ith, i ∈ [1, N ]sub]Average equivalent RCS of the flight path of the strips;
(10)F10: sub track limit distance difference
Figure BDA0002367446100000051
Wherein,
Figure BDA0002367446100000052
represents the ith, i ∈ [1, N ]sub]Difference of limit distance, r, of strip flight pathl (i),1≤l≤LsubRepresents the ith, i ∈ [1, N ]sub]The distance between the first track point in the strip track and the receiving station;
(11)F11: sub track smoothness
Figure BDA0002367446100000053
Wherein, Smooth(i)Represents the ith, i ∈ [1, N ]sub]The smoothness of the flight path of the strip,
Figure BDA0002367446100000054
represents the ith, i ∈ [1, N ]sub]The distance from the (l + 1) th track point in the strip track to the connecting line of the (l + 2) th track point;
(12)F12: degree of sub track focus
Figure BDA0002367446100000055
Wherein, focus(i)Represents the ith, i ∈ [1, N ]sub]The degree of focusing of the flight path of the strip,
Figure BDA0002367446100000056
Figure BDA0002367446100000057
Figure BDA0002367446100000058
s.t, (objectto) represents a constraint;
wherein,
Figure BDA0002367446100000059
is the X-axis distance representing the ith track point
Figure BDA00023674461000000510
Distance from the X axis of the c track point
Figure BDA00023674461000000511
The difference between the difference of the two phases,
Figure BDA00023674461000000512
y-axis distance representing the ith track point
Figure BDA00023674461000000513
Distance from the c track point Y axis
Figure BDA00023674461000000514
The difference between the two;
(13)F13: sub track position entropy
Figure BDA00023674461000000515
Wherein,
Figure BDA00023674461000000516
represents the ith, i ∈ [1, N ]sub]Position entropy of the strip flight path;
(14)F14: sub-track RCS entropy
Figure BDA0002367446100000061
Wherein,
Figure BDA0002367446100000062
represents the ith, i ∈ [1, N ]sub]RCS entropy of strip track.
In one embodiment, S2 includes:
s2.1: collecting original track data of various types of targets with comparison information;
s2.2: integrating all the original track data of the same type of targets collected in the S2.1 into a long original track, and extracting sub-tracks from the long original track;
s2.3: the characteristic quantities F of the objects to be classified in the sub-tracks extracted from S2.2 and S1m,m∈[1,M]A typical value of (a), wherein Ci,i∈[1,I]Characteristic quantity F corresponding to type objectm,m∈[1,M]Typical values of (1) include the minimum value
Figure BDA0002367446100000063
Mean value of
Figure BDA0002367446100000064
Maximum value
Figure BDA0002367446100000065
The specific manner of each typical value is as follows:
Figure BDA0002367446100000066
Figure BDA0002367446100000067
Figure BDA0002367446100000068
wherein,
Figure BDA0002367446100000069
represents from Ci,i∈[1,I]And the total number of the sub tracks extracted from all the original tracks of the type target.
In one embodiment, S4 specifically includes:
s4.1: obtaining the target type C, C e { C ∈ { C1,C2,...,CIFeature set of
Figure BDA00023674461000000610
Is a characteristic quantity FkSetting a decision section [ F ]k,min,Fk,max]；
S4.2: obtaining characteristic quantity F of each sub-track of target to be classifiedkWherein the ith, i ∈ [1, N ∈ ]sub]Characteristic quantity F of strip flight pathkIs expressed as
Figure BDA00023674461000000611
The following judgment function is set:
Figure BDA00023674461000000612
wherein, Delta epsilon (0, 1)]A threshold, func (F), which is a predetermined proportion of the number of sub-tracksk) Is 0 or 1 when func (F)k) When the target track to be classified is 1, judging that the target track to be classified belongs to a target type C, wherein C belongs to { C ∈ { C }1,C2,...,CI}; when func (F)k) When the target track to be classified is 0, judging that the target track to be classified does not belong to a target type C, wherein C belongs to { C ∈ }1,C2,...,CI}；
S4.3: obtaining a judgment function func (F) corresponding to each characteristic quantity according to the measured value of the characteristic quantity corresponding to each target typek1),...,func(Fk) And according to the judgment function func (F) corresponding to each characteristic quantityk1),...,func(Fk) The final judgment result is the original track which is screened out and belongs to the target type corresponding to the target typical characteristic quantity, and the target type corresponding to the target typical characteristic quantity is used as the type of the track.
In one embodiment, S4.3 specifically includes:
s4.3.1: when the judgment function result corresponding to each feature quantity is selected to be processed by the AND operation, if func (F)k1)&&...&&func(Fk) If the target track to be classified belongs to the target type C, judging that C belongs to { C ∈ [ C ]1,C2,...,CIIf func (F)k1)&...&func(Fk) If the target track to be classified is 0, judging that the target track to be classified does not belong to a target type C, wherein C belongs to { C ∈ [ C ]1,C2,...,CITherein of&&Represents an and operation;
s4.3.2: when the judgment function result corresponding to each characteristic quantity is selected to be processed by adopting OR operation, if func (F)k1)||...||func(Fk) If the target track to be classified belongs to the target type C, judging that C belongs to { C ∈ [ C ]1,C2,...,CIIf func (F)k1)||...||func(Fk) If the target track to be classified is 0, judging that the target track to be classified does not belong to a target type C, wherein C belongs to { C ∈ [ C ]1,C2,...,CIAnd (c) wherein | | | represents an or operation.
In one embodiment, S5 specifically includes:
s5.1: according to typical values of each characteristic quantity of each type of target obtained by statistics from actually measured flight path data, the characteristic quantity is Fm,m∈[1,M]Assigning a weight value wm,m∈[1,M]And satisfy
Figure BDA0002367446100000071
The larger the difference of the different types of targets in the expression of a certain characteristic quantity is, the larger the weight value distributed by the characteristic quantity is;
s5.2: obtaining each sub track pair weight value wm,m∈[1,M]Has a contribution of wm/Nsub,m∈[1,M]Setting the utilization characteristic quantity Fm,m∈[1,M]Judging whether the object belongs to Ci,i∈[1,I]Has a confidence of
Figure BDA0002367446100000072
Then
Figure BDA0002367446100000073
The calculation method of (2) is as follows:
Figure BDA0002367446100000074
wherein,
Figure BDA0002367446100000081
wherein, Fm,nCharacteristic quantity F of nth sub-track representing target track to be classifiedm,m∈[1,M]The measured value of (a) is measured,
Figure BDA0002367446100000082
features F of various types of targets obtained by utilizing measured flight path statistics of various types of targetsm,m∈[1,M]The minimum, average and maximum of the typical values of (c); judging whether the target track to be detected belongs to the type C by utilizing all the characteristicsi,i∈[1,I]Can be expressed as
Figure BDA0002367446100000083
S5.3: if for all j e [1, I]And j ≠ i, all have
Figure BDA0002367446100000084
Then the object to be classified is judged to be CiA type object.
One or more technical solutions in the embodiments of the present application have at least one or more of the following technical effects:
the invention provides a classification and identification method of an external radiation source radar target based on track feature extraction, which comprises the steps of firstly obtaining an original track of a target to be classified, extracting a sub-track from the original track of the target to be classified, and calculating each feature quantity of the target to be classified according to the extracted sub-track; then, counting the actual measurement track data to obtain typical values of each characteristic quantity of each type of target; secondly, determining the typical characteristic quantity of the target for rough classification according to the typical value of each characteristic quantity of each type of target; and then, according to the determined target typical characteristic quantity, carrying out rough classification on the original flight path of the target to be classified, then according to the relation between each characteristic value of the remaining target flight path to be classified and the typical value of each characteristic quantity, calculating the confidence coefficient of the remaining target flight path to be classified corresponding to each type of target, and according to the confidence coefficient, carrying out fine classification on the remaining target flight path to be classified.
The method provided by the invention is used for operating the target track points, has small calculated amount, can well integrate the prior information of the characteristics of various types of targets (namely the typical values of various characteristic quantities of each type of targets obtained by statistics in the actually measured track data) into the classification and identification process, firstly carries out coarse classification on the original tracks of the targets to be classified according to the determined typical characteristic quantities of the targets, further calculates the confidence degrees of the residual tracks to be classified corresponding to various types of targets, and carries out fine classification on the residual tracks to be classified according to the confidence degrees, thereby not only improving the accuracy of the classification and identification of the targets, but also ensuring the robustness of the algorithm, and the classification and identification results have very clear physical significance and have interpretability.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
FIG. 1: the invention provides a flow chart of an external radiation source radar target classification and identification method based on track feature extraction.
FIG. 2: the flight path diagram of 3 types of flight paths collected in the embodiment of the invention.
FIG. 3: the invention is a schematic diagram of the acquisition of the sub-track.
FIG. 4: the navigation path of the navigation airplane obtained in the 'typical path rough classification' process in the embodiment of the invention.
FIG. 5: the results of 3 types of flight path classification in the embodiment of the invention.
FIG. 6: the navigation aircraft track map extracted after classification and identification in the embodiment of the invention is obtained.
Detailed Description
On the basis of analyzing the advantages and the disadvantages of the existing radar target classification and identification method in detail, the invention aims to fully mine target track characteristic information and provide an external radiation source radar target classification and identification method which is small in calculation amount and easy to realize in engineering.
In order to achieve the above object, the main concept of the present invention is as follows:
the method for classifying and identifying the targets of the radar with the external radiation source based on the track feature extraction constructs feature quantities including aspects of motion states, echo features, track expression forms and the like, and fully excavates the feature differences of different types of target tracks. The method comprises the steps of firstly extracting sub tracks from original tracks to be classified, calculating values of all characteristic quantities of the sub tracks, then analyzing various target tracks of known types collected in advance, obtaining typical ranges of all characteristic quantities corresponding to all types of targets (typical values of all characteristic quantities of all types of targets are obtained through statistics from actually measured track data), and finally calculating confidence degrees of the tracks to be classified corresponding to all types of targets according to the value conditions of all characteristic quantities of the tracks to be classified to realize target classification and identification. Each step of the method has clear physical meaning, good interpretability, small calculated amount, easy engineering realization and popularization and application value.
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The embodiment provides a classification and identification method for an external radiation source radar target based on track feature extraction, please refer to fig. 1, and the method includes:
s1: the method comprises the steps of obtaining an original flight path of a target to be classified, extracting a sub-flight path from the original flight path of the target to be classified, and calculating various characteristic quantities of the target to be classified according to the extracted sub-flight path.
Specifically, the feature quantities of the target to be classified proposed in S1 include a plurality of feature quantities, which are extracted and calculated from the sub-tracks and used for subsequent classification and identification.
S2: counting typical values of each characteristic quantity of each type of target from actually measured track data;
specifically, the measured track data refers to known type of track data, that is, priori data, and is used to obtain typical values of each feature quantity through statistics.
It should be noted that the execution sequence of S1 and S2 may be adjusted, and S1 may be executed first, and then S2 may be executed, or S2 may be executed first, and then S1 may be executed, or the execution may be executed simultaneously.
S3: and determining the typical characteristic quantities of the targets for rough classification according to the typical values of the characteristic quantities of each type of target, wherein the difference between the typical value of the typical characteristic quantity of the target on the first type of target and the typical values of other types of targets on the typical characteristic quantity of the target is larger than a preset value, and the first type of target and the other types of targets are different types.
Specifically, S3 is to determine a target characteristic feature quantity that can be used for classification according to the typical value of each feature quantity, specifically, it can be determined whether there is a difference between the typical value of some feature quantities of some types of targets and the typical value of other types of targets on the feature quantity, which is greater than a preset value, that is, there is a significant difference (there is little or no overlap with the overlap interval of other types of targets on the feature quantity), if there is no such typical feature quantity, then the target characteristic feature quantity is not used for rough classification, and the process proceeds directly to S5 for final classification recognition; if so, the operation of S4 is performed
S4: and carrying out rough classification on the original flight path of the target to be classified according to the determined target typical characteristic quantity, screening out the original flight path of the target type corresponding to the target typical characteristic quantity, and taking the original flight path of the target type not corresponding to the target typical characteristic quantity as the residual target flight path to be classified.
Specifically, after the target characteristic feature quantity is determined, the original flight path of the target to be classified can be roughly classified by using the target characteristic feature quantity.
S5: and calculating the confidence degrees of the residual target tracks to be classified corresponding to various types of targets according to the relation between each characteristic value of the residual target tracks to be classified and the typical value of each characteristic quantity, and classifying the residual target tracks to be classified according to the confidence degrees.
Specifically, the tracks that cannot be classified in S4, that is, the tracks of the targets to be classified, are classified by the confidence in S5.
The implementation of the present invention is described below with reference to an embodiment implemented in an external radiation source radar system based on digital television signals (DTMB). The radar receiver is located near a suburb airport in Luoyang city, Henan province and is used for monitoring the navigation aircraft trained in an airspace near the airport in real time. Besides the navigation aircraft, the radar also detects other moving targets including birds and ground vehicles at the same time in the monitoring process, and forms a track similar to the navigation aircraft on the radar display terminal, so that the judgment of a radar operator on the navigation aircraft is greatly influenced. The present embodiment is directed to distinguishing the category attribute of each flight path on a radar display terminal by using the target classification recognition algorithm provided in the present invention. In this embodiment, 1439 tracks of different types of targets are collected, including 300 tracks of navigable aircraft (with ADSB information comparison), 540 tracks of birds (without comparison information, possibly mixed with few non-bird tracks), 599 tracks of unknown targets (without comparison information, possibly including various targets such as birds, ground vehicles, navigable aircraft, civil aircraft, and other clutter). 30% of each type of flight path is randomly extracted for analyzing the condition of each characteristic of each type of flight path (training process), and the rest 70% of the flight path is used for testing the effect of classification recognition. Namely, the track conditions of each type of track training and verification are shown in the following table:
TABLE 1 training and validation of track numbers
Navigation aircraft track Bird track Unknown track
Number of training tracks 90 162 180
Verifying track number 210 378 419
The radar system with the external radiation source is provided with a reference antenna pointing to a transmitting station and used for receiving direct wave signals (reference signals), and a monitoring antenna array pointing to a radar observation area and used for receiving target echo signals (monitoring signals). After signal processing processes such as reference signal reconstruction, direct wave and multipath clutter suppression, two-dimensional cross correlation, target detection, parameter estimation and the like are carried out, tracking processing is carried out on a detection point track to obtain a target track, and the obtained target track is processed by the method provided by the invention.
In one embodiment, S1 specifically includes:
s1.1: acquiring an original track of a target to be classified;
s1.2: carrying out sliding windowing operation on the original track of the target to be classified, forming track points contained in each sliding window into sub tracks of the original track, and extracting the sub tracks;
s1.3: calculating each characteristic quantity of the target to be classified from the extracted sub-tracks, wherein the mth characteristic quantity is represented as Fm,m∈[1,M]Then F ism,m∈[1,M]Comprises the following steps:
(1)F1: average distance of sub-tracks
Figure BDA0002367446100000121
Wherein L issubRepresenting the length of each sub-track, NsubThe number of word tracks which can be extracted from the original track with the length L is represented,
Figure BDA0002367446100000122
represents the ith, i ∈ [1, N ]sub]Average distance of flight path of strip, rl (i)Represents the ith, i ∈ [1, N ]sub]The L, L in the flight path of the strip belongs to [1, L ∈ ]sub]The distance of each waypoint from the receiving station;
(2)F2: sub track average velocity
Figure BDA0002367446100000123
Wherein,
Figure BDA0002367446100000124
represents the ith, i ∈ [1, N ]sub]The average speed of the flight path of the strip,
Figure BDA0002367446100000125
represents the ith, i ∈ [1, N ]sub]The L, L in the flight path of the strip belongs to [1, L ∈ ]sub]The speed of each track point;
(3)F3: sub track average acceleration
Figure BDA0002367446100000126
Wherein,
Figure BDA0002367446100000127
represents the ith, i ∈ [1, N ]sub]The average speed of the flight path of the strip,
Figure BDA0002367446100000128
represents the ith, i ∈ [1, N ]sub]The L, L in the flight path of the strip belongs to [1, L ∈ ]sub]The speed of each track point;
(4)F4: average height of sub-track
Figure BDA0002367446100000129
Wherein,
Figure BDA00023674461000001210
represents the ith, i ∈ [1, N ]sub]The average height of the flight path of the strip,
Figure BDA00023674461000001211
represents the ith, i ∈ [1, N ]sub]The L, L in the flight path of the strip belongs to [1, L ∈ ]sub]The height of each course point;
(5)F5: sub-track target equivalent mean RCS: cross section of radar scattering
Figure BDA0002367446100000131
Wherein,
Figure BDA0002367446100000132
represents the ith, i ∈ [1, N ]sub]The average equivalent RCS of the strip flight path,
Figure BDA0002367446100000133
represents the ith, i ∈ [1, N ]sub]The L, L in the flight path of the strip belongs to [1, L ∈ ]sub]The equivalent RCS of each track point, A represents a constant related to system noise coefficient, signal transmitting power, receiving and transmitting antenna gain and other parameters, RrIndicating the distance, R, of the target from the receiving radar stationtRepresents the distance of the target from the transmitting station (which can be obtained by the double-base distance sum of the target), lambda represents the wavelength of the transmitted signal, and SNR represents the signal-to-noise ratio of the target;
(6)F6: sub track velocity variance
Figure BDA0002367446100000134
Wherein,
Figure BDA0002367446100000135
represents the ith, i ∈ [1, N ]sub]The variance of the speed of the flight path of the strip,
Figure BDA0002367446100000136
represents the ith, i ∈ [1, N ]sub]The L, L in the flight path of the strip belongs to [1, L ∈ ]sub]The speed of each of the track points is,
Figure BDA0002367446100000137
represents the ith, i ∈ [1, N ]sub]Average speed of the flight path of the sliver;
(7)F7: sub track acceleration variance
Figure BDA0002367446100000138
Wherein,
Figure BDA0002367446100000139
represents the ith, i ∈ [1, N ]sub]The variance of the acceleration of the strip flight path,
Figure BDA00023674461000001310
represents the ith, i ∈ [1, N ]sub]The L, L in the flight path of the strip belongs to [1, L ∈ ]sub]The acceleration of each of the waypoints is,
Figure BDA00023674461000001311
represents the ith, i ∈ [1, N ]sub]Average acceleration of the strip flight path;
(8)F8: sub track height variance
Figure BDA00023674461000001312
Wherein,
Figure BDA00023674461000001313
represents the ith, i ∈ [1, N ]sub]The variance of the height of the flight path of the strip,
Figure BDA00023674461000001314
represents the ith, i ∈ [1, N ]sub]The L, L in the flight path of the strip belongs to [1, L ∈ ]sub]The height of each course point is determined,
Figure BDA00023674461000001315
represents the ith, i ∈ [1, N ]sub]Average height of the flight path;
(9)F9: sub-track equivalent RCS variance
Figure BDA0002367446100000141
Wherein,
Figure BDA0002367446100000142
represents the ith, i ∈ [1, N ]sub]The equivalent RCS of the strip flight path,
Figure BDA0002367446100000143
represents the ith, i ∈ [1, N ]sub]The L, L in the flight path of the strip belongs to [1, L ∈ ]sub]The equivalent RCS of each track point,
Figure BDA0002367446100000144
represents the ith, i ∈ [1, N ]sub]Average equivalent RCS of the flight path of the strips;
(10)F10: sub track limit distance difference
Figure BDA0002367446100000145
Wherein,
Figure BDA0002367446100000146
represents the ith, i ∈ [1, N ]sub]Difference of limit distance, r, of strip flight pathl (i),1≤l≤LsubRepresents the ith, i ∈ [1, N ]sub]The distance between the first track point in the strip track and the receiving station;
(11)F11: sub track smoothness
Figure BDA0002367446100000147
Wherein, Smooth(i)Represents the ith, i ∈ [1, N ]sub]The smoothness of the flight path of the strip,
Figure BDA0002367446100000148
represents the ith, i ∈ [1, N ]sub]The distance from the (l + 1) th track point in the strip track to the connecting line of the (l + 2) th track point;
(12)F12: degree of sub track focus
Figure BDA0002367446100000149
Wherein, focus(i)Represents the ith, i ∈ [1, N ]sub]The degree of focusing of the flight path of the strip,
Figure BDA00023674461000001410
Figure BDA00023674461000001411
Figure BDA00023674461000001412
s.t, (objectto) represents a constraint;
wherein,
Figure BDA00023674461000001413
is the X-axis distance representing the ith track point
Figure BDA00023674461000001414
Distance from the X axis of the c track point
Figure BDA00023674461000001415
The difference between the difference of the two phases,
Figure BDA00023674461000001416
y-axis distance representing the ith track point
Figure BDA00023674461000001417
Distance from the c track point Y axis
Figure BDA00023674461000001418
The difference between the two;
(13)F13: sub track position entropy
Figure BDA0002367446100000151
Wherein,
Figure BDA0002367446100000152
represents the ith, i ∈ [1, N ]sub]Position entropy of the strip flight path;
(14)F14: sub-track RCS entropy
Figure BDA0002367446100000153
Wherein,
Figure BDA0002367446100000154
represents the ith, i ∈ [1, N ]sub]RCS entropy of strip track.
Specifically, in step 1The original track has a length of L, that is, the original track includes L track points, and each track point includes K-dimensional feature information, so that the original track information can be represented as an L × K matrix X ═ X1 x2 ... xN]TIn the present invention, the ith, i ∈ [1, L ]]A track point xi,i∈[1,L]The following information is contained: x-axis distance rx,iY-axis distance ry,iX axis velocity vx,iVelocity v of y-axisy,iX-axis acceleration ax,iY-axis acceleration ay,iHeight hiPower piSNR of signal to noise ratioiDistance of double base and rb,iVelocity v of double baseb,iSignal frequency fi；
The sub track in S2 is obtained by sliding windowing the original track in S1, and the track points included in each window constitute the sub track of the original track. Setting the length of each sub track as LsubThe length of overlap between adjacent sub-tracks is LoThen the original track with length L can be obtained
Figure BDA0002367446100000155
Strip track, where floor (. cndot.) means rounding down, i.e. discarding the insufficient L at the end of the original tracksubPart of the track points.
Ith, i ∈ [1, N ]sub]The strip track can be expressed as
Figure BDA0002367446100000156
As shown in fig. 2, which is a schematic diagram of a sub-track acquiring method, in this embodiment, the length L of each sub-track is setsub10, overlap length L between adjacent sub-tracks o5. A schematic diagram of the neutron path is shown in fig. 3.
S3, extracting each item of characteristic information used for subsequent classification and identification in the sub track; setting M kinds of feature quantities to be extracted for subsequent classification and identification, and representing the mth feature quantity as Fm,m∈[1,M]Then F ism,m∈[1,M]Including several types of characteristic quantities mentioned above.
The above characteristic quantities can be selected according to actual conditions, and can include partial quantities and not necessarily all the quantitiesAnd (4) a section. For example, when the radar system of the external radiation source does not have the height measuring capability, so F4And F8The features do not participate in the subsequent object classification process.
In one embodiment, S2 includes:
s2.1: collecting original track data of various types of targets with comparison information;
s2.2: integrating all the original track data of the same type of targets collected in the S2.1 into a long original track, and extracting sub-tracks from the long original track;
s2.3: the characteristic quantities F of the objects to be classified in the sub-tracks extracted from S2.2 and S1m,m∈[1,M]A typical value of (a), wherein Ci,i∈[1,I]Characteristic quantity F corresponding to type objectm,m∈[1,M]Typical values of (1) include the minimum value
Figure BDA0002367446100000161
Mean value of
Figure BDA0002367446100000162
Maximum value
Figure BDA0002367446100000163
The specific manner of each typical value is as follows:
Figure BDA0002367446100000164
Figure BDA0002367446100000165
Figure BDA0002367446100000166
wherein,
Figure BDA0002367446100000167
represents from Ci,i∈[1,I]All originals of type objectAnd extracting the total number of the obtained sub tracks from the tracks.
Specifically, typical values of each characteristic of each type of target are obtained through statistics from actually measured flight path data, classification of flight paths with obvious partial characteristics is achieved through the typical values (the process is called as 'coarse flight path classification based on typical characteristics'), and original flight path data of various types of targets with comparison information are collected in S2.1, wherein the original flight path data can be civil aviation and navigation aircraft targets with ADSB comparison information, unmanned aerial vehicle targets with GPS information comparison, vehicle targets with road information comparison, and the like. For example, let C be collected1,C2,...,CIThe original flight paths of the I types of targets are shared; in the embodiment, 3 different types of tracks are collected together, that is, in the embodiment, I is 3, the track of the navigable aircraft has ADSB comparison information, and the other two types of tracks have no comparison information;
s2.2, integrating all the original tracks of the same type of targets collected in the step S2.1 into a long original track, and extracting sub-tracks in a sliding windowing mode in the step S1.2;
s2.3 obtaining the characteristic quantities F calculated in S1 from the S2.2 sub-tracksm,m∈[1,M]Typical values of (a) for subsequent object classification recognition, Ci,i∈[1,I]Characteristic quantity F corresponding to type objectm,m∈[1,M]Typical values of (1) include the minimum value
Figure BDA0002367446100000168
Mean value of
Figure BDA0002367446100000169
Maximum value
Figure BDA00023674461000001610
In one embodiment, S4 specifically includes:
s4.1: obtaining the target type C, C e { C ∈ { C1,C2,...,CIFeature set of
Figure BDA0002367446100000171
Is a characteristic quantity FkSetting a decision section [ F ]k,min,Fk,max]；
S4.2: obtaining characteristic quantity F of each sub-track of target to be classifiedkWherein the ith, i ∈ [1, N ∈ ]sub]Characteristic quantity F of strip flight pathkIs expressed as
Figure BDA0002367446100000172
The following judgment function is set:
Figure BDA0002367446100000173
wherein, Delta epsilon (0, 1)]A threshold, func (F), which is a predetermined proportion of the number of sub-tracksk) Is 0 or 1 when func (F)k) When the target track to be classified is 1, judging that the target track to be classified belongs to a target type C, wherein C belongs to { C ∈ { C }1,C2,...,CI}; when func (F)k) When the target track to be classified is 0, judging that the target track to be classified does not belong to a target type C, wherein C belongs to { C ∈ }1,C2,...,CI}；
S4.3: obtaining a judgment function func (F) corresponding to each characteristic quantity according to the measured value of the characteristic quantity corresponding to each target typek1),...,func(Fk) And according to the judgment function func (F) corresponding to each characteristic quantityk1),...,func(Fk) The final judgment result is the original track which is screened out and belongs to the target type corresponding to the target typical characteristic quantity, and the target type corresponding to the target typical characteristic quantity is used as the type of the track.
Specifically, when there is a target characteristic quantity, let the type target be C, C e { C ∈ { C }1,C2,...,CIJudging whether the target track to be classified belongs to a target type C according to the following steps, wherein C belongs to { C ∈ }1,C2,...,CIJudging, wherein C represents a determined type. Set object type C, C e { C1,C2,...,CIFeature set of
Figure BDA0002367446100000174
Is used for type judgment of the target track to be detected, and takes a feature set { Fk1,...,FkCharacteristic quantity F ink,Fk∈{Fk1,...,FkExplaining a specific judgment process by taking an example as the characteristic quantity FkSetting a decision section [ F ]k,min,Fk,max]；
Setting the original flight path of the target to be classified and extracting N according to the method in S1.2subStrip flight path and according to the characteristics F of S1.3kThe characteristic quantity F of each sub-track of the target to be classified is obtainedkIs set as the ith, i is belonged to [1, N [ ]sub]Characteristic quantity F of strip flight pathkIs expressed as
Figure BDA0002367446100000175
The aforementioned decision function is set.
For example, if the original flight path is to be classified into three categories, civil aircraft, unmanned aerial vehicle, and bird. By adopting the method, the data are analyzed to find that the two characteristic quantities of the speed and the height of the civil aircraft are greatly different from those of the unmanned aerial vehicle and the bird. At this time, the speed and the altitude are selected as typical characteristic quantities for judging whether the target belongs to the civil aircraft (in this case, C represents the civil aircraft), and then a range can be set for the speed and the altitude, for example, the speed range is set to be 100 m/s-200 m/s, the altitude is set to be 2000 m-3000 m, and if the speed and the altitude of one flight path are in the range, the target is considered to be the civil aircraft by the invention.
The invention aims to classify the original flight path, so that the unit of the flight path classification is the original flight path. For example, there is an original flight path, 10 sub-flight paths can be extracted from the original flight path, then the method of the present invention is adopted to extract features from the sub-flight paths, and then the features of the 10 sub-flight paths are integrated to judge which category the original flight path belongs to finally. The sub-track can be regarded as a means, and the attribute (category) of the original track is judged by using the extracted features of the sub-track.
In one embodiment, S4.3 specifically includes:
s4.3.1: when the judgment function result corresponding to each feature quantity is selected to be processed by the AND operation, if func (F)k1)&&...&&func(Fk) If the target track to be classified belongs to the target type C, judging that C belongs to { C ∈ [ C ]1,C2,...,CIIf func (F)k1)&...&func(Fk) If the target track to be classified is 0, judging that the target track to be classified does not belong to a target type C, wherein C belongs to { C ∈ [ C ]1,C2,...,CITherein of&&Represents an and operation;
s4.3.2: when the judgment function result corresponding to each characteristic quantity is selected to be processed by adopting OR operation, if func (F)k1)||...||func(Fk) If the target track to be classified belongs to the target type C, judging that C belongs to { C ∈ [ C ]1,C2,...,CIIf func (F)k1)||...||func(Fk) If the target track to be classified is 0, judging that the target track to be classified does not belong to a target type C, wherein C belongs to { C ∈ [ C ]1,C2,...,CIAnd (c) wherein | | | represents an or operation.
Specifically, when the object type C is comprehensively utilized, C e { C1,C2,...,CIFeature set of { F }k1,...,FkWhen classifying and identifying the target to be classified, calculating each characteristic quantity according to the steps 4.1 and 4.2 to obtain a judgment function func (F) corresponding to each characteristic quantityk1),...,func(Fk) And (3) selecting the final judgment result according to the actual measurement test condition and the and operation or the or operation for each judgment function result, as described in steps S4.3.1 and S4.3.2.
In this embodiment, the feature quantity { F is set by analyzing the condition of each feature value obtained in S2.31,F2,F5The mean distance, the mean speed and the mean equivalent RCS of the sub-tracks respectively correspond to are taken as the rough classification characteristic of the typical navigation aircraft track, and the characteristic thresholds corresponding to the 3 characteristic quantities are respectively [ F [1,min,F1.max]＝[15000,40000]、[F2,min,F2.max]＝[40,80][F3,min,F3.max]＝[155,180]Setting these three characteristic quantities simultaneouslySub-track number ratio threshold delta is [0.3,0.3 ═]And the three characteristics adopt an AND operation mode, and the physical meaning of the rough classification of the typical navigation aircraft tracks is that when the average distance of 30 percent of sub-tracks simultaneously satisfied in one track is in [15000,40000 ]]The average speed of 30% of the sub-tracks in the m range is [40,80 ]]In the m/s range, the equivalent average RCS of 30% of the sub-tracks is [155,180 ]]If the range is within the range, the flight path is determined to be the flight path of the navigable aircraft, in this embodiment, through the rough classification process described in step 4, the attributes of 42 flight paths are determined, and the accuracy is 100%, where fig. 4 is a navigable aircraft flight path diagram determined in this step.
The remaining 965 routes that were not classified and identified are sent to step 5 for final classification and identification.
In one embodiment, S5 specifically includes:
s5.1: according to typical values of each characteristic quantity of each type of target obtained by statistics from actually measured flight path data, the characteristic quantity is Fm,m∈[1,M]Assigning a weight value wm,m∈[1,M]And satisfy
Figure BDA0002367446100000191
The larger the difference of the different types of targets in the expression of a certain characteristic quantity is, the larger the weight value distributed by the characteristic quantity is;
s5.2: obtaining each sub track pair weight value wm,m∈[1,M]Has a contribution of wm/Nsub,m∈[1,M]Setting the utilization characteristic quantity Fm,m∈[1,M]Judging whether the object belongs to Ci,i∈[1,I]Has a confidence of
Figure BDA0002367446100000192
Then
Figure BDA0002367446100000193
The calculation method of (2) is as follows:
Figure BDA0002367446100000194
wherein,
Figure BDA0002367446100000195
wherein, Fm,nCharacteristic quantity F of nth sub-track representing target track to be classifiedm,m∈[1,M]The measured value of (a) is measured,
Figure BDA0002367446100000196
features F of various types of targets obtained by utilizing measured flight path statistics of various types of targetsm,m∈[1,M]The minimum, average and maximum of the typical values of (c); judging whether the target track to be detected belongs to the type C by utilizing all the characteristicsi,i∈[1,I]Can be expressed as
Figure BDA0002367446100000201
S5.3: if for all j e [1, I]And j ≠ i, all have
Figure BDA0002367446100000202
Then the object to be classified is judged to be CiA type object.
Specifically, in S5.1, [ w ] is set based on the results of analyzing the characteristics of the three types of tracks1,w2,w3,w5,w6,w7,w9,w10,w11,w12]＝[0.1,0.2,0.1,0.1,0.1,0.05,0.05,0.1,0.1,0.1]Since the radar system of the external radiation source used by the invention does not have the height measuring capability, F in 1.34And F8Is not used in the object classification recognition process, so that F does not need to be set4And F8Corresponding weight value w4And w8。
Then, a confusion matrix of the classification recognition result of the embodiment is calculated, as shown in fig. 5, it can be seen that the recognition accuracy of the navigable aircraft is 91.4%, the recognition accuracy of the bird is 83.6%, the recognition accuracy of the unknown track is 84.2%, and the total recognition accuracy of all tracks is 85.5%. The misclassification situation between the bird track and the unknown track in the classification result is serious, and the main reason may be that neither of the two types of tracks has comparison information, so that whether the bird track contains a small part of other types of tracks or not can not be completely determined, and the unknown track also may contain a part of bird tracks, so that the characteristics of the two types of tracks are overlapped greatly, and the classification recognition effect is poor. As shown in fig. 6, the three types of flight paths are extracted by the algorithm provided by the present invention, and obviously, compared with the flight path situation before the classification in fig. 2, the present invention can effectively extract the flight path of the navigable aircraft, and has a great application potential.
The specific embodiments described herein are merely illustrative of the spirit of the invention. Various modifications or additions may be made to the described embodiments or alternatives may be employed by those skilled in the art without departing from the spirit or ambit of the invention as defined in the appended claims.

Claims (5)

1. An external radiation source radar target classification and identification method based on track feature extraction is characterized by comprising the following steps:
s1: acquiring an original flight path of a target to be classified, extracting a sub-flight path from the original flight path of the target to be classified, and calculating each characteristic quantity of the target to be classified according to the extracted sub-flight path;
s2: counting typical values of each characteristic quantity of each type of target from actually measured track data;
s3: determining a target typical characteristic quantity for rough classification according to typical values of various characteristic quantities of each type of target, wherein the difference between the typical value of the target typical characteristic quantity on the first type of target and the typical values of other types of targets on the target typical characteristic quantity is larger than a preset value, and the first type of target and the other types of targets are different types;
s4: roughly classifying the original flight paths of the targets to be classified according to the determined target typical characteristic quantities, screening out original flight paths of target types corresponding to the target typical characteristic quantities, and taking the original flight paths of the target types not corresponding to the target typical characteristic quantities as the residual target flight paths to be classified;
s5: calculating the confidence of the remaining target tracks to be classified corresponding to various types of targets according to the relation between each characteristic value of the remaining target tracks to be classified and the typical value of each characteristic quantity, and classifying the remaining target tracks to be classified according to the confidence;
wherein, S1 specifically includes:
s1.1: acquiring an original track of a target to be classified;
s1.2: carrying out sliding windowing operation on the original track of the target to be classified, forming track points contained in each sliding window into sub tracks of the original track, and extracting the sub tracks;
s1.3: calculating each characteristic quantity of the target to be classified from the extracted sub-tracks, wherein the mth characteristic quantity is represented as Fm,m∈[1,M]Then F ism,m∈[1,M]Comprises the following steps:
(1)F1: average distance of sub-tracks
Figure FDA0003237836890000011
Wherein L issubRepresenting the length of each sub-track, NsubThe number of sub-tracks that can be extracted from the original track with the length L is represented,
Figure FDA0003237836890000012
represents the ith, i ∈ [1, N ]sub]Average distance of flight path of strip, rl (i)Represents the ith, i ∈ [1, N ]sub]The L, L in the flight path of the strip belongs to [1, L ∈ ]sub]The distance of each waypoint from the receiving station;
(2)F2: sub track average velocity
Figure FDA0003237836890000021
Wherein,
Figure FDA0003237836890000022
represents the ith, i ∈ [1, N ]sub]The average speed of the flight path of the strip,
Figure FDA0003237836890000023
represents the ith, i ∈ [1, N ]sub]The L, L in the flight path of the strip belongs to [1, L ∈ ]sub]The speed of each track point;
(3)F3: sub track average acceleration
Figure FDA0003237836890000024
Wherein,
Figure FDA0003237836890000025
represents the ith, i ∈ [1, N ]sub]The average acceleration of the flight path of the strip,
Figure FDA0003237836890000026
represents the ith, i ∈ [1, N ]sub]The L, L in the flight path of the strip belongs to [1, L ∈ ]sub]Acceleration of each course point;
(4)F4: average height of sub-track
Figure FDA0003237836890000027
Wherein,
Figure FDA0003237836890000028
represents the ith, i ∈ [1, N ]sub]The average height of the flight path of the strip,
Figure FDA0003237836890000029
represents the ith, i ∈ [1, N ]sub]The L, L in the flight path of the strip belongs to [1, L ∈ ]sub]The height of each course point;
(5)F5: sub-track target equivalent mean RCS: cross section of radar scattering
Figure FDA00032378368900000210
Wherein,
Figure FDA00032378368900000211
represents the ith, i ∈ [1, N ]sub]The average equivalent RCS of the strip flight path,
Figure FDA00032378368900000212
represents the ith, i ∈ [1, N ]sub]The L, L in the flight path of the strip belongs to [1, L ∈ ]sub]Equivalent RCS of each track point, A represents a constant related to a system noise coefficient, signal transmission power, and a receiving/transmitting antenna gain parameter, RrIndicating the distance, R, of the target from the receiving radar stationtRepresenting the distance of the target from the transmitting station, λ representing the wavelength of the transmitted signal, and SNR representing the target signal-to-noise ratio;
(6)F6: sub track velocity variance
Figure FDA00032378368900000213
Wherein,
Figure FDA00032378368900000214
represents the ith, i ∈ [1, N ]sub]The speed variance of the strip flight path;
(7)F7: sub track acceleration variance
Figure FDA0003237836890000031
Wherein,
Figure FDA0003237836890000032
represents the ith, i ∈ [1, N ]sub]Acceleration variance of the strip flight path;
(8)F8: sub track height variance
Figure FDA0003237836890000033
Wherein,
Figure FDA0003237836890000034
represents the ith, i ∈ [1, N ]sub]Height variance of the strip flight path;
(9)F9: sub-track equivalent RCS variance
Figure FDA0003237836890000035
Wherein,
Figure FDA0003237836890000036
represents the ith, i ∈ [1, N ]sub]Equivalent RCS of the strip flight path;
(10)F10: sub track limit distance difference
Figure FDA0003237836890000037
Wherein,
Figure FDA0003237836890000038
represents the ith, i ∈ [1, N ]sub]The difference of the limit distance of the strip flight path;
(11)F11: sub track smoothness
Figure FDA0003237836890000039
Wherein, Smooth(i)Represents the ith, i ∈ [1, N ]sub]The smoothness of the flight path of the strip,
Figure FDA00032378368900000310
represents the ith, i ∈ [1, N ]sub]The distance from the (l + 1) th track point in the strip track to the connecting line of the (l + 2) th track point;
(12)F12: degree of sub track focus
Figure FDA00032378368900000311
Wherein, focus(i)Represents the ith, i ∈ [1, N ]sub]The degree of focusing of the flight path of the strip,
Figure FDA00032378368900000312
Figure FDA00032378368900000313
Figure FDA0003237836890000041
s.t, (subject to) represents a constraint;
wherein,
Figure FDA0003237836890000042
is the X-axis distance representing the ith track point
Figure FDA0003237836890000043
Distance from the X axis of the c track point
Figure FDA0003237836890000044
The difference between the difference of the two phases,
Figure FDA0003237836890000045
y-axis distance representing the ith track point
Figure FDA0003237836890000046
Distance from the c track point Y axis
Figure FDA0003237836890000047
The difference between the two;
(13)F13: sub track position entropy
Figure FDA0003237836890000048
Wherein,
Figure FDA0003237836890000049
represents the ith, i ∈ [1, N ]sub]Position entropy of the strip flight path;
(14)F14: sub-track RCS entropy
Figure FDA00032378368900000410
Wherein,
Figure FDA00032378368900000411
represents the ith, i ∈ [1, N ]sub]RCS entropy of strip track.
2. The method of claim 1, wherein S2 includes:
s2.1: collecting original track data of various types of targets with comparison information;
s2.2: integrating all the original track data of the same type of targets collected in the S2.1 into a long original track, and extracting sub-tracks from the long original track;
s2.3: seed extracted from S2.2Characteristic quantities F of objects to be classified in flight path and S1m,m∈[1,M]A typical value of (a), wherein Ci,i∈[1,I]Characteristic quantity F corresponding to type objectm,m∈[1,M]Typical values of (1) include the minimum value
Figure FDA00032378368900000412
Mean value of
Figure FDA00032378368900000413
Maximum value
Figure FDA00032378368900000414
The specific manner of each typical value is as follows:
Figure FDA00032378368900000415
Figure FDA00032378368900000416
Figure FDA00032378368900000417
wherein,
Figure FDA00032378368900000418
represents from Ci,i∈[1,I]And the total number of the sub tracks extracted from all the original tracks of the type target.
3. The method of claim 1, wherein S4 specifically comprises:
s4.1: obtaining the target type C, C e { C ∈ { C1,C2,...,CIFeature set of
Figure FDA0003237836890000051
Is a characteristic quantity FkSetting a decision section [ F ]k,min,Fk,max]；
S4.2: obtaining characteristic quantity F of each sub-track of target to be classifiedkWherein the ith, i ∈ [1, N ∈ ]sub]Characteristic quantity F of strip flight pathkIs expressed as
Figure FDA0003237836890000052
The following judgment function is set:
Figure FDA0003237836890000053
wherein, Delta epsilon (0, 1)]A threshold, func (F), which is a predetermined proportion of the number of sub-tracksk) Is 0 or 1 when func (F)k) When the target track to be classified is 1, judging that the target track to be classified belongs to a target type C, wherein C belongs to { C ∈ { C }1,C2,...,CI}; when func (F)k) When the target track to be classified is 0, judging that the target track to be classified does not belong to a target type C, wherein C belongs to { C ∈ }1,C2,...,CI}；
S4.3: obtaining a judgment function func (F) corresponding to each characteristic quantity according to the measured value of the characteristic quantity corresponding to each target typek1),...,func(Fk) And according to the judgment function func (F) corresponding to each characteristic quantityk1),...,func(Fk) The final judgment result is the original track which is screened out and belongs to the target type corresponding to the target typical characteristic quantity, and the target type corresponding to the target typical characteristic quantity is used as the type of the track.
4. The method according to claim 3, wherein S4.3 specifically comprises:
s4.3.1: when the judgment function result corresponding to each feature quantity is selected to be processed by the AND operation, if func (F)k1)&&...&&func(Fk) If the target track to be classified belongs to the target type C, judging that C belongs to { C ∈ [ C ]1,C2,...,CIIf func (F)k1)&...&func(Fk) If the target track to be classified is 0, judging that the target track to be classified does not belong to a target type C, wherein C belongs to { C ∈ [ C ]1,C2,...,CITherein of&&Represents an and operation;
s4.3.2: when the judgment function result corresponding to each characteristic quantity is selected to be processed by adopting OR operation, if func (F)k1)||...||func(Fk) If the target track to be classified belongs to the target type C, judging that C belongs to { C ∈ [ C ]1,C2,...,CIIf func (F)k1)||...||func(Fk) If the target track to be classified is 0, judging that the target track to be classified does not belong to a target type C, wherein C belongs to { C ∈ [ C ]1,C2,...,CIAnd (c) wherein | | | represents an or operation.
5. The method of claim 1, wherein S5 specifically comprises:
s5.1: according to typical values of each characteristic quantity of each type of target obtained by statistics from actually measured flight path data, the characteristic quantity is Fm,m∈[1,M]Assigning a weight value wm,m∈[1,M]And satisfy
Figure FDA0003237836890000054
The larger the difference of the different types of targets in the expression of a certain characteristic quantity is, the larger the weight value distributed by the characteristic quantity is;
s5.2: obtaining each sub track pair weight value wm,m∈[1,M]Has a contribution of wm/Nsub,m∈[1,M]Setting the utilization characteristic quantity Fm,m∈[1,M]Judging whether the object belongs to Ci,i∈[1,I]Has a confidence of
Figure FDA0003237836890000061
Then
Figure FDA0003237836890000062
The calculation method of (2) is as follows:
Figure FDA0003237836890000063
wherein,
Figure FDA0003237836890000064
wherein, Fm,nCharacteristic quantity F of nth sub-track representing target track to be classifiedm,m∈[1,M]The measured value of (a) is measured,
Figure FDA0003237836890000065
features F of various types of targets obtained by utilizing measured flight path statistics of various types of targetsm,m∈[1,M]The minimum, average and maximum of the typical values of (c); judging whether the target track to be detected belongs to the type C by utilizing all the characteristicsi,i∈[1,I]The confidence of (d) is expressed as:
Figure FDA0003237836890000066
s5.3: if for all j e [1, I]And j ≠ i, all have
Figure FDA0003237836890000067
Then the object to be classified is judged to be CiA type object.
CN202010040144.5A 2020-01-15 2020-01-15 External radiation source radar target classification and identification method based on track feature extraction Active CN111142085B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010040144.5A CN111142085B (en) 2020-01-15 2020-01-15 External radiation source radar target classification and identification method based on track feature extraction

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010040144.5A CN111142085B (en) 2020-01-15 2020-01-15 External radiation source radar target classification and identification method based on track feature extraction

Publications (2)

Publication Number Publication Date
CN111142085A CN111142085A (en) 2020-05-12
CN111142085B true CN111142085B (en) 2021-12-03

Family

ID=70525012

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010040144.5A Active CN111142085B (en) 2020-01-15 2020-01-15 External radiation source radar target classification and identification method based on track feature extraction

Country Status (1)

Country Link
CN (1) CN111142085B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111860688B (en) * 2020-07-31 2023-10-31 北京无线电测量研究所 Target type identification method and system and electronic equipment
CN113640767B (en) * 2021-08-13 2023-11-14 北京理工大学 Variance-based low-resolution radar target identification method
CN113900093B (en) * 2021-09-07 2024-04-26 武汉大学 Automatic target recognition method for external radiation source radar based on diversity technology
CN113687348B (en) * 2021-09-18 2024-03-15 无锡威孚高科技集团股份有限公司 Pedestrian recognition method and device based on tracking micro Doppler graph

Family Cites Families (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104133211B (en) * 2014-07-07 2016-08-24 中国船舶重工集团公司第七二四研究所 A kind of Doppler frequency transradar target classification identification method
CN104076354B (en) * 2014-07-08 2017-01-18 西安电子科技大学 Detection method for radar target tracks on basis of correlation speeds
CN104502906B (en) * 2014-12-09 2017-02-01 中国民航大学 Spatial ultrahigh-speed maneuvered target detection method based on RMDCFT (Radon-Modified Discrete Chirp-Fourier Transform)
CN106405537B (en) * 2016-08-23 2018-11-16 西安电子科技大学 Radar track initial mode based on location information and doppler information
CN107688170B (en) * 2017-08-21 2020-08-28 哈尔滨工业大学 Radar target track starting method based on random forest
CN109685138B (en) * 2018-12-25 2023-04-07 东南大学 XLPE power cable partial discharge type identification method
CN110045362B (en) * 2019-03-01 2022-02-22 合肥工业大学 RNN-based radar target track identification method and device
CN110208789B (en) * 2019-05-29 2021-04-13 西安空间无线电技术研究所 Multi-frame combined target track identification method
CN110427986B (en) * 2019-07-16 2022-02-01 浙江大学 Target classification method of kernel support vector machine based on millimeter wave radar point cloud characteristics
CN110533695A (en) * 2019-09-04 2019-12-03 深圳市唯特视科技有限公司 A kind of trajectory predictions device and method based on DS evidence theory

Also Published As

Publication number Publication date
CN111142085A (en) 2020-05-12

Similar Documents

Publication Publication Date Title
CN111142085B (en) External radiation source radar target classification and identification method based on track feature extraction
Stumpf et al. The National Severe Storms Laboratory mesocyclone detection algorithm for the WSR-88D
CN107238821B (en) Airport runway foreign matter detection method and device based on characteristic spectrum characteristics
EP3167305B1 (en) A classification system for radar and sonar applications
CN109409225B (en) Unmanned aerial vehicle classification method and device based on radar multipath signal time-frequency feature fusion
PL220849B1 (en) Weather and airborne clutter suppression using a cluster shape classifier
CN109901130B (en) Rotor unmanned aerial vehicle detection and identification method based on Radon transformation and improved 2DPCA
CN107506794B (en) Decision tree-based ground moving object classification algorithm
CN111401168B (en) Multilayer radar feature extraction and selection method for unmanned aerial vehicle
CN107247262B (en) A kind of airfield runway foreign matter layer detection method
CN109117776B (en) Aircraft and meteorological clutter classification and identification method based on flight path information
Chen et al. Classification of UAV and bird target in low-altitude airspace with surveillance radar data
CN109753874A (en) A kind of low slow small classification of radar targets method based on machine learning
CN114139373A (en) Multi-sensor automatic cooperative management method for unmanned aerial vehicle reverse braking
CN107133555B (en) Method for identifying 8-character motion track target
Zong-ling et al. Remote sensing ship target detection and recognition system based on machine learning
Sinha et al. Estimation of Doppler profile using multiparameter cost function method
CN111796250A (en) False trace point multi-dimensional hierarchical suppression method based on risk assessment
Kida et al. Improvement of rain/no-rain classification methods for microwave radiometer observations over the ocean using a 37 GHz emission signature
CN112213703B (en) Method and device for inverting cloud parameters by millimeter wave cloud radar
Cohn et al. Radial velocity and wind measurement with NIMA–NWCA: Comparisons with human estimation and aircraft measurements
Ahmad et al. Tracking of target body and micro-Doppler components in drone surveillance radar
Han et al. Aircraft target detection in polsar image based on region segmentation and multi-feature decision
CN113625266A (en) Method, device, storage medium and equipment for detecting low-speed target by using radar
CN113537347B (en) Unmanned aerial vehicle and flying bird target classification method based on track motion characteristics

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