CN109117776A - Aircraft and meteorological clutter classifying identification method based on track information - Google Patents
Aircraft and meteorological clutter classifying identification method based on track information Download PDFInfo
- Publication number
- CN109117776A CN109117776A CN201810868550.3A CN201810868550A CN109117776A CN 109117776 A CN109117776 A CN 109117776A CN 201810868550 A CN201810868550 A CN 201810868550A CN 109117776 A CN109117776 A CN 109117776A
- Authority
- CN
- China
- Prior art keywords
- feature
- aircraft
- layer
- sequence
- meteorological clutter
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Signal Processing (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The present invention relates to a kind of aircrafts based on track information and meteorological clutter classifying identification method, first according to the difference of aircraft and meteorological clutter track information change, extract the feature that can reflect target intrinsic propesties;Then apparent feature is distinguished using aircraft and meteorological clutter, part sample, propulsion layer-by-layer so finally carry out Classification and Identification to remaining sample using classifier, complete the Classification and Identification to all aircrafts and meteorological clutter out for identification.For solving the problems, such as that aircraft and meteorological clutter track are multifarious.
Description
Technical field
The invention belongs to Radar Target Recognition field, can be used for miscellaneous to the Aircraft Targets and meteorology that form track
Wave carries out Classification and Identification.
Background technique
When radar detection aerial target, various angels are frequently present of around target, one of the most common is exactly gas
As clutter, when meteorological clutter has certain movement velocity, and can form track, the echo very phase of they and Aircraft Targets
Seemingly, it can not usually be inhibited by traditional filtering method, just need to be distinguish them using the means of classification at this time.
In the low resolution of radar, the size of target is less than the resolution ratio of radar, and the echo of target, which is one, to be had
The point of amplitude and phase, information contained by single echo are less.Therefore the identification for moving target, common means are to utilize
The information in multiple echo periods, i.e. the micro-doppler information of target, but it is (higher to obtain accurate target micro-doppler information
DOPPLER RESOLUTION), it usually needs longer residence time, but for search radar, the function of search of radar is wished
Direction in space as much as possible is irradiated within the limited time, this residence time that may cause target on each direction will not
It is very long or even very short, in this case, the fine motion information that each target significantly can not can be used for mutually distinguishing.
It can use the information that target is repeatedly searched for for situation above in order to obtain target information as much as possible
Variation carries out Classification and Identification, namely carries out Classification and Identification to the target for forming track, to improve the accuracy of Classification and Identification.Shape
Mainly have at the information that each mark includes in the target of track: amplitude, distance, orientation, pitching, scanning to time, linear velocity,
Height etc..
To the classified uses of Aircraft Targets and meteorological clutter be their track information changes difference, therefore in order to realize
Classification to Aircraft Targets and meteorological clutter needs the variation to their track information to carry out feature extraction, and extraction can reflect mesh
The feature for marking intrinsic propesties, makes the difference of similar target become smaller, the difference of heterogeneous destinations becomes larger.Feature extraction is usually one and incites somebody to action
Classification accuracy rate can be improved in the process that original echoed signals are mapped from higher dimensional space to lower dimensional space, effective feature,
Shorten the classification time.
In order to extract the feature that can effectively identify Aircraft Targets and meteorological clutter, need to observe the change of their track information
Change.The much information that a wherein typical track for Aircraft Targets and meteorological clutter is set forth in Fig. 1-Fig. 6 compares, including returns
One changes amplitude (carrying out normalizing using distance), height, azimuth, consecutive points mark trace interval, linear velocity, acceleration etc..
In practical applications, it is often impossible to just classify to entire track after building boat completely, it thus can be in thunder
There is too many track on up to display screen, and major part therein all may not be true Aircraft Targets, thus lose
Go the purpose of classification, it is therefore desirable to just classify at the initial stage for building boat, once meet the needs of confidence level, so that it may directly
The track for being judged to meteorological clutter is removed, the track of display aircraft.If the point mark number of every track is N, i=1 ..., N are indicated
The vector of 1~i mark information compositions is then set as a sample, it would be desirable to each by i-th mark in track
Sample carries out Classification and Identification, therefore the sample number of every track is N-i+1.
For Aircraft Targets and meteorological clutter, their track changes multiplicity, in order to be obtained preferably with the smallest operand
Classification results, this method using layering identification strategy.
Summary of the invention
Technical problems to be solved
In order to avoid the shortcomings of the prior art, the present invention proposes a kind of aircraft and meteorological clutter based on track information
Classifying identification method passes through when Aircraft Targets and meteorological clutter can be considerably less with fine motion information and extracts aircraft and meteorological clutter
Track information change feature, and using the methods of layering identification, complete to their Classification and Identification.
Technical solution
A kind of aircraft based on track information and meteorological clutter classifying identification method, it is characterised in that steps are as follows:
Step 1: according to the difference of aircraft and meteorological clutter track information change, extracting the spy that can reflect target intrinsic propesties
Sign:
Feature 1: the maximum value of amplitude sequence is normalized
The expression formula of the normalization amplitude of target:
Snr_normilize=20*log10 (SNR)+40*log10 (R/r)
Wherein, SNR is signal-to-noise ratio, and R is target range, and normalized distance is wanted in r expression;It is same due in different positions
The normalization amplitude of target can change, therefore the maximum value for normalizing amplitude sequence indicates in the Multiple-Scan of target most
Big amplitude;
Feature 2: it normalizes the average value of amplitude sequence: being indicated using the average value of normalization amplitude sequence;
Feature 3: the normalization variance of amplitude sequence is normalized
Wherein:Indicate normalization amplitude sequence,
Feature 4: the mean value of high degree of sequence: the statistical value of target flight height;
Feature 5: the mean value of consecutive points mark trace interval sequence: using the equal of consecutive points mark trace interval sequence
Value characterizes the statistics trace interval of aircraft and meteorological clutter;
Feature 6: the number in consecutive points mark trace interval sequence greater than thresholding T accounts for the ratio of entire sequence length;
Feature 7: the mean value of linear speed degree series: the statistical value of target flight speed;
Feature 8: the number in consecutive points trace speed difference sequence equal to 0 accounts for the ratio of entire sequence length;
Feature 9: the variation range of azimuth sequence;
Step 2: layering identification:
First layer: civil aviaton is identified using feature 1 and feature 7;
The second layer: to the remaining sample of first layer, part meteorological clutter is identified using feature 4;
Third layer: to the remaining sample of the second layer, part Aircraft Targets are identified using feature 3;
4th layer: to the remaining sample of third layer, identifying part Aircraft Targets using feature 5 and feature 8;
Layer 5: to the 4th layer of remaining sample, part meteorological clutter is identified using feature 5 and feature 8;
Layer 6: to the remaining sample of layer 5, part Aircraft Targets are identified using feature 1;
Layer 7: Classification and Identification is carried out to the remaining sample of layer 6 using classifier, which has automated characterization
The ability of selection, the feature selected are 1,2,3,6,8,9.
Classification and Identification is carried out using linear classifier feature selecting Method Using Relevance Vector Machine FSRVM in step 2.
Beneficial effect
A kind of aircraft based on track information proposed by the present invention and meteorological clutter classifying identification method, have the beneficial effect that
1) present invention is suitable for solving that the Classification and Identification of aircraft and meteorological clutter of the fine motion information when less can be used;2) present invention is according to winged
The difference of machine and meteorological clutter track information change is extracted the feature that can effectively distinguish them;3) present invention is known using layering
Other strategy, for solving the problems, such as that aircraft and meteorological clutter track are multifarious.
Detailed description of the invention
Fig. 1 compares for the normalization amplitude of aircraft and meteorological clutter;
Fig. 2 compares for the azimuth of aircraft and meteorological clutter;
Fig. 3 compares for the height of aircraft and meteorological clutter;
Fig. 4 compares for the linear velocity of aircraft and meteorological clutter;
Fig. 5 compares for the consecutive points mark trace interval of aircraft and meteorological clutter;
Fig. 6 be aircraft and meteorological clutter acceleration ratio compared with;
Fig. 7 compares for the consecutive points trace speed difference of aircraft and meteorological clutter;
Fig. 8 is the track schematic diagram of civil aviaton;
Fig. 9 is the track schematic diagram of unmanned plane;
Figure 10 is the track schematic diagram of meteorology 1;
Figure 11 is the track schematic diagram of meteorology 2;
Figure 12 is the two dimensional character distribution (feature 1 and 7) of aircraft and meteorological clutter;
Figure 13 is the one-dimensional characteristic distribution (feature 4) of aircraft and meteorological clutter;
Figure 14 is the one-dimensional characteristic distribution (feature 3) of aircraft and meteorological clutter;
Figure 15 is the two dimensional character distribution (feature 5 and 8) of aircraft and meteorological clutter;
Figure 16 is the two dimensional character distribution (feature 5 and 8) of aircraft and meteorological clutter;
Figure 17 is the one-dimensional characteristic distribution (feature 1) of aircraft and meteorological clutter;
Figure 18 is the variation with the variation discrimination of sample midpoint mark number;
Figure 19 is with the increase for participating in fusion number of samples, and discrimination changes after unmanned plane fusion.
Specific embodiment
Now in conjunction with embodiment, attached drawing, the invention will be further described:
Aircraft based on track information and meteorological Clutter Classification method the following steps are included:
S1: according to the difference of aircraft and meteorological clutter track information change, the feature that can reflect target intrinsic propesties is extracted.
The feature of extraction mainly has: 1) normalizing the maximum value of amplitude sequence;2) average value of amplitude sequence is normalized;3) width is normalized
The normalization variance of degree series;4) mean value of high degree of sequence;5) mean value of consecutive points mark trace interval sequence;6) consecutive points
Number in mark trace interval sequence greater than thresholding T accounts for the ratio of entire sequence length;7) mean value of linear speed degree series;8)
Number in consecutive points trace speed difference sequence equal to 0 accounts for the ratio of entire sequence length;9) the variation model of azimuth sequence
It encloses;10) accelerate the entropy of degree series.
S2: in order to preferably classify to aircraft and meteorological clutter, method for distinguishing is known using layering.The principle of layering
It is: distinguishes apparent feature first with aircraft and meteorological clutter, for identification part sample, so successively propulsion out, finally
Classification and Identification is carried out to remaining sample using classifier, completes the Classification and Identification to all aircrafts and meteorological clutter.
It is of the invention that the specific implementation steps are as follows:
S1: according to the difference of aircraft and meteorological clutter track information change, the feature that can reflect target intrinsic propesties is extracted.
It is described as follows:
By the comparative analysis to much information in aircraft in Fig. 1-Fig. 6 and meteorological clutter track, it is extracted following several
Feature:
Feature 1: the maximum value of amplitude sequence is normalized
From figure 1 it appears that the normalization amplitude of aircraft and meteorological clutter has biggish difference, therefore it can use this
Feature is classified, and the solution of normalization amplitude is given below.
Radar equation is as shown in (1) formula:
The calculation formula that the radar cross section (RCS) of target is derived according to radar equation, such as formula (2):
Wherein M is pulse accumulation number;R is target range;δ is target cross section;PtFor transmission power;τ is transmitting pulse
Width;G is antenna gain;λ is wavelength;K is Boltzmann constant;T0For temperature;FnFor receiver noise factor;SNR is noise
Than;L is loss coefficient caused by radar is lost.From formula (2) as can be seen that when radar system parameters determine, in addition to target
Distance and signal-to-noise ratio except, other variables are constant or definite value, therefore for the target detected, it is only necessary to know that mesh
Target distance and signal-to-noise ratio can be obtained by the normalization amplitude of target:
Snr_normilize=20*log10 (SNR)+40*log10 (R/r) (3)
Normalized distance is wanted in wherein r expression.
Since in different positions, the normalization amplitude of same target can change, therefore normalize amplitude sequence most
Big value indicates the maximum amplitude in the Multiple-Scan of target.
Feature 2: the average value of amplitude sequence is normalized
The statistics of same target normalizes amplitude in order to obtain, is indicated here with the average value of normalization amplitude sequence.
Feature 3: the normalization variance of amplitude sequence is normalized
From figure 1 it appears that aircraft is different with the fluctuation characteristic of the normalization amplitude of meteorological clutter, and variance can be very
The fluctuation characteristic of good characterization sequence, therefore extract feature of the variance of normalization amplitude sequence as classification.The general of variance is determined
Justice are as follows:
Wherein:Indicate normalization amplitude sequence,
The variance obtained due to this mode there are amplitude sensitive needs variance to be normalized, it may be assumed that
Feature 4: the mean value of high degree of sequence
What the mean value of high degree of sequence characterized is the statistical value of target flight height.The flying height of different target is usually not
With, such as civil aviaton generally fly it is relatively high, and unmanned plane generally fly it is relatively low, the height of different cloud layers is not also identical, because
This feature can be used in combination with other feature.
Feature 5: the mean value of consecutive points mark trace interval sequence
From figure 5 it can be seen that the trace interval between Aircraft Targets and the most of consecutive points marks of meteorological clutter exists
4000 (this is the scan period of radar system) left and right, and in 98 marks of Aircraft Targets, between only one consecutive points mark
Sweep time is greater than 5000, and meteorological clutter has the trace interval between more consecutive points mark to be greater than 5000, this explanation exists
In scanning each time, aircraft can be seen substantially, and meteorological clutter is due to the unstability of itself, it can not be each time
It can see in scanning, therefore trace interval is longer.Therefore we use consecutive points mark trace interval sequence here
Mean value characterize the statistics trace interval of aircraft and meteorological clutter.
Feature 6: the number in consecutive points mark trace interval sequence greater than thresholding T accounts for the ratio of entire sequence length
Similar to feature 5, this feature also for characterization aircraft and meteorological clutter consecutive points mark trace interval not
Together, it is only illustrated from another angle.Here the setting of thresholding T is counted according to specific data, than
It is greater than 4000 if the setting of T in Fig. 5 and less than 8000.
Feature 7: the mean value of linear speed degree series
What the mean value of linear speed degree series characterized is the statistical value of target flight speed.The flying speed of different target is usually
Different, for example the flying speed of civil aviaton, fighter plane is usually relatively high, and the flying speed of meteorological clutter generally will not be too
Height, therefore can use this feature to characterize their difference.
Feature 8: the number in consecutive points trace speed difference sequence equal to 0 accounts for the ratio of entire sequence length
Figure 4, it is seen that the linear velocity fluctuations of aircraft and meteorological clutter have biggish difference, aircraft is most of
The movement that time remains a constant speed, and meteorological clutter, due to unstability, the big rise and fall of linear velocity, it is gentle that Fig. 7 gives aircraft
As the comparison of clutter consecutive points trace speed difference, therefrom it can be seen that, Aircraft Targets have a more zero, and meteorological clutter zero
Value is very few, therefore the ratio of entire sequence length can be accounted for the number for being equal to 0 in consecutive points trace speed difference sequence
To characterize their difference.
Feature 9: the variation range of azimuth sequence
From figure 2 it can be seen that Aircraft Targets are since mobility is stronger, for example its flight orientation can be left-to-right.So
Afterwards and from right to left, namely in the presence of the case where turning around of turning round, the azimuth range that will have Aircraft Targets at this time compares
Small situation, and the traffic direction of meteorological clutter in a short time can be generally consistent, therefore its azimuthal transformation model
It encloses general bigger.
Feature 10: accelerate the entropy of degree series
Entropy is probabilistic for describing being averaged for information source, therefore entropy is an effectively feature.It can be with from Fig. 6
Find out, the acceleration profile of Aircraft Targets compares concentration, and the acceleration profile of meteorological clutter is more dispersed, therefore can use
The entropy of acceleration characterizes their difference.If the sequence of acceleration isacnThe probability of appearance is pn, then acceleration
Sequence entropy is defined as:
Wherein:The dispersion level of acceleration is characterized with entropy, stroll is more concentrated, and entropy is smaller.
S2: layering identification.It is described as follows:
The data that the present invention uses be certain ground search radar acquisition measured data, there are two types of Aircraft Targets: civil aviaton and
Unmanned plane, wherein there are 272 tracks in civil aviaton, and unmanned plane has 140 tracks;Meteorological clutter is acquired in different time different location
Data, indicated respectively with meteorology 1 and meteorology 2, wherein meteorology 1 has 285 tracks, meteorology 2 has 456 tracks, forms boat
The point mark number of mark differs.Fig. 8~Figure 11 gives the schematic diagram of aircraft and meteorological noise component track.It can from figure
Out, meteorological clutter is distributed in entire radar asorbing paint plane, therefore removes meteorological clutter and be highly desirable.
In order to obtain preferable classification performance with less operation time, before mentioned the present invention will using layering know
The principle of other strategy, layering is exactly: extraction Aircraft Targets and meteorological clutter first distinguish apparent feature, are used to identification division
Then sample recycles other feature successively to promote, is finally completed the classification as clutter gentle to all aircraft class targets.Using
Layering identification structure can reduce the complex distribution of each subclass target signature, can be special for different object optimum choices
Sign, improves the utilization rate of feature, and system complexity is also reduced while improving discrimination, therefore is layered same in identification
When there is the processes of feature selecting.In addition, being directed to the radar, need to ensure while Aircraft Targets not false dismissal, as far as possible
Reduce the false alarm rate of meteorological clutter.The present invention is below described in detail each layer using seven layers of identification.
First layer: civil aviaton is identified using feature 1 and feature 7
For aircarrier aircraft, usually it normalization amplitude it is bigger, and flying speed is relatively high, therefore can use
The two features identify civil aviaton, and Figure 12 gives the bidimensional characteristic pattern of all sample characteristics 1 and 7, and wherein red line is setting
Thresholding, it can be seen from the figure that feature distribution meets our analysis, it, can be completely civil aviaton and portion using the two features
Unmanned plane is divided to identify.The sample number of initial aircraft is 47819, and the sample number of meteorological clutter is 32544.
The second layer: to the remaining sample of first layer, part meteorological clutter is identified using feature 4
In first layer, identifies that flying speed is bigger or normalized the biggish aircraft of amplitude, flown remaining
In machine namely speed is smaller and the normalization lesser aircraft of amplitude, and typically flivver, flying height will not be too high, and gas
It as clutter, such as cloud, then may move in high-altitude, therefore can use elevation information and identify part meteorological clutter, such as scheme
Shown in 13,4 distribution schematic diagram of feature of remaining sample after first layer classification is given, it can be seen from the figure that part is meteorological miscellaneous
The height of wave is greater than the height of unmanned plane, therefore thresholding is arranged and classifies, as shown in red line in figure.
Third layer: to the remaining sample of the second layer, part Aircraft Targets are identified using feature 3
Aircraft to have been analyzed in a upper section and meteorological normalization amplitude scintillation is different, the posture one of aircraft changes,
Corresponding normalization amplitude will change namely the posture of aircraft is more sensitive, therefore can use feature normalization amplitude
The normalization variance of sequence is to aircraft and meteorological progress Classification and Identification.Figure 14 gives the distribution of third layer residue sample characteristics 3
Schematic diagram, wherein red line indicates the thresholding of setting, it can be seen from the figure that aircraft and this meteorological feature difference are more bright
It is aobvious.
4th layer: to the remaining sample of third layer, identifying part Aircraft Targets using feature 5 and feature 8
Figure 15 gives the two dimensional character distribution schematic diagram of third layer residue sample characteristics 5 and feature 8, can from figure
Out, just to feature 5 as the analysis of feature 8, Aircraft Targets generally all compare a section as above relative to meteorological clutter feature 5
It is small, and feature 8 is general all bigger, therefore thresholding can be set and identify part Aircraft Targets, as shown in red line in figure.
Layer 5: to the 4th layer of remaining sample, part meteorological clutter is identified using feature 5 and feature 8
Similar to the 4th layer, we equally can use the meteorological clutter that feature 5 and 8 identifies part, because they
Feature 5 is general all bigger, and feature 8 is general all smaller, and as shown in figure 16, red line is classification thresholding in figure.
Layer 6: to the remaining sample of layer 5, part Aircraft Targets are identified using feature 1
By analyzing the remaining sample of layer 5, find in remaining sample at this time, the part in Aircraft Targets
The normalization amplitude of sample is bigger, as shown in figure 17, therefore the Aircraft Targets that thresholding identifies part is arranged, such as red line in figure
It is shown.
Layer 7: Classification and Identification is carried out to the remaining sample of layer 6 using classifier
In order to reduce engineering operation amount, this layer we mainly utilize linear classifier feature selecting Method Using Relevance Vector Machine
(Feature Selection with Relevance Vector Machine, FSRVM) carries out Classification and Identification, the classifier
Ability with automated characterization selection.Aircraft sample has 1786 in this layer of remaining sample, and meteorological sample has 6186, therefrom
Uniformly extract training sample: 358, aircraft, 442 meteorological, the feature selected is 1,2,3,6,8,9.
By the details for each layering being recited above, it is given below using layering recognition strategy to all samples
Final discrimination: the discrimination 0.9918 (sample number 47819) of Aircraft Targets, the discrimination of meteorological clutter: 0.9447 (sample number
32544)。
In engineering, we are more concerned at the initial stage for forming track, the discrimination of each sample, because to each track,
Our processing method be constantly accumulate the process for forming sample as step-length using mark since the first mark, therefore under
Face provides the variation of the variation discrimination with sample midpoint mark number, as shown in figure 18, in order to increase the confidence level of identification, this
In set each sample midpoint mark number and be at least 8 (in practical application, minimum suitable mark number of different features can be with
Physical features, a mark samples such as difference, such as normalization amplitude, height may obtain classification results, common practice
It is self-adaptive processing, once target property meet demand, so that it may judgement is made, is to show with 8 here for convenience of explanation
Example), wherein each corresponding test sample number of point mark may be different, because the point mark number in every track is different
, when it is less than the point mark number of setting, which is just not involved in test.It can be seen from the figure that forming the first of track
Begin, the discrimination of Aircraft Targets and meteorological target is all 95% or so, with an increase for mark number, the discrimination of Aircraft Targets
Higher and higher, the discrimination of meteorological clutter is declined, but still be greater than 85%, and this result also comply with we setting
Guarantee while Aircraft Targets not false dismissal, as far as possible the principle of the false alarm rate of reduction meteorological clutter.
The recognition result for the only single sample that Figure 18 is provided, for each track, we can use multisample and melt
The method of conjunction increases discrimination, and the technology of fusion can effectively improve the recognition performance of target, and the method for fusion has very much, most simply
Be ballot method, basic principle is " the minority is subordinate to the majority ".It is illustrated by taking three samples fusions as an example below, table 1 gives
Civil aviaton, unmanned plane, meteorology 1 and meteorology 2 sample be respectively a point mark 1~8, discrimination when 1~9,1~10 and they three
A obtained fusion recognition rate of voting.As can be seen from the table, fused discrimination is not highest, this is because participating in
The sample of ballot is less, only there are three, even if in this way, fusion result be also second high, be to a certain extent compromise
Selection should be using which as final result because being not aware that in three samples in practical applications.Figure 19 gives
Unmanned plane is with the increase for participating in fusion number of samples, the situation of change of discrimination after fusion, it can be seen from the figure that with melting
The increase of number of samples is closed, discrimination gradually increases.
The gentle discrimination and fused discrimination as when sample point mark number is respectively 8,9,10 of 1 aircraft of table
Claims (2)
1. a kind of aircraft based on track information and meteorological clutter classifying identification method, it is characterised in that steps are as follows:
Step 1: according to the difference of aircraft and meteorological clutter track information change, extract the feature that can reflect target intrinsic propesties:
Feature 1: the maximum value of amplitude sequence is normalized
The expression formula of the normalization amplitude of target:
Snr_normilize=20*log10 (SNR)+40*log10 (R/r)
Wherein, SNR is signal-to-noise ratio, and R is target range, and normalized distance is wanted in r expression;Due in different positions, same target
Normalization amplitude can change, therefore normalize amplitude sequence maximum value indicate it is maximum in the Multiple-Scan of target
Amplitude;
Feature 2: it normalizes the average value of amplitude sequence: being indicated using the average value of normalization amplitude sequence;
Feature 3: the normalization variance of amplitude sequence is normalized
Wherein:Indicate normalization amplitude sequence,
Feature 4: the mean value of high degree of sequence: the statistical value of target flight height;
Feature 5: the mean value of consecutive points mark trace interval sequence: using the mean value of consecutive points mark trace interval sequence come
Characterize the statistics trace interval of aircraft and meteorological clutter;
Feature 6: the number in consecutive points mark trace interval sequence greater than thresholding T accounts for the ratio of entire sequence length;
Feature 7: the mean value of linear speed degree series: the statistical value of target flight speed;
Feature 8: the number in consecutive points trace speed difference sequence equal to 0 accounts for the ratio of entire sequence length;
Feature 9: the variation range of azimuth sequence;
Step 2: layering identification:
First layer: civil aviaton is identified using feature 1 and feature 7;
The second layer: to the remaining sample of first layer, part meteorological clutter is identified using feature 4;
Third layer: to the remaining sample of the second layer, part Aircraft Targets are identified using feature 3;
4th layer: to the remaining sample of third layer, identifying part Aircraft Targets using feature 5 and feature 8;
Layer 5: to the 4th layer of remaining sample, part meteorological clutter is identified using feature 5 and feature 8;
Layer 6: to the remaining sample of layer 5, part Aircraft Targets are identified using feature 1;
Layer 7: Classification and Identification is carried out to the remaining sample of layer 6 using classifier, which selects with automated characterization
Ability, the feature selected be 1,2,3,6,8,9.
2. a kind of aircraft based on track information according to claim 1 and meteorological clutter classifying identification method, feature
It is in step 2 to carry out Classification and Identification using linear classifier feature selecting Method Using Relevance Vector Machine FSRVM.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810868550.3A CN109117776B (en) | 2018-08-02 | 2018-08-02 | Aircraft and meteorological clutter classification and identification method based on flight path information |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810868550.3A CN109117776B (en) | 2018-08-02 | 2018-08-02 | Aircraft and meteorological clutter classification and identification method based on flight path information |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109117776A true CN109117776A (en) | 2019-01-01 |
CN109117776B CN109117776B (en) | 2021-09-07 |
Family
ID=64851607
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810868550.3A Active CN109117776B (en) | 2018-08-02 | 2018-08-02 | Aircraft and meteorological clutter classification and identification method based on flight path information |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109117776B (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109782247A (en) * | 2019-01-28 | 2019-05-21 | 中船重工鹏力(南京)大气海洋信息系统有限公司 | A method of utilizing track Information revision radar return |
CN110261837A (en) * | 2019-06-27 | 2019-09-20 | 中国航空工业集团公司雷华电子技术研究所 | A kind of complex target RCS calculation method based on track information |
CN110501683A (en) * | 2019-08-19 | 2019-11-26 | 杭州电子科技大学 | A kind of extra large land Clutter Classification method based on 4 D data feature |
CN111458701A (en) * | 2020-04-12 | 2020-07-28 | 西安电子工程研究所 | Meteorological track restraining method based on track characteristic iterative updating |
CN112001342A (en) * | 2020-08-28 | 2020-11-27 | 电子科技大学 | Clutter classification method adopting VGG-16 network |
Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20070024494A1 (en) * | 2005-04-11 | 2007-02-01 | Dizaji Reza M | Classification system for radar and sonar applications |
EP2420972A1 (en) * | 2010-08-19 | 2012-02-22 | BAE SYSTEMS plc | Sensor data processing |
KR20120060359A (en) * | 2010-12-02 | 2012-06-12 | 한국항공우주산업 주식회사 | Method of recognizing flight pattern of rotor craft, apparatus for recognizing flight pattern implementing the same and recording medium |
CN103885043A (en) * | 2014-03-31 | 2014-06-25 | 西安电子科技大学 | Airplane target clutter and noise stable classification method based on generalized matched filtering |
CN105242254A (en) * | 2015-10-22 | 2016-01-13 | 中国船舶重工集团公司第七二四研究所 | Air target identification method based on data quality assessment |
CN106093935A (en) * | 2016-08-26 | 2016-11-09 | 西安电子科技大学 | Pitching widebeam radar ground and low target sorting technique |
CN106483513A (en) * | 2016-09-20 | 2017-03-08 | 西安电子工程研究所 | A kind of aircraft class target micro-doppler texture characteristic extracting method |
CN107132515A (en) * | 2017-05-09 | 2017-09-05 | 哈尔滨工业大学 | A kind of point mark screening technique constrained based on multidimensional information |
CN107145850A (en) * | 2017-04-28 | 2017-09-08 | 西安电子科技大学 | A kind of target identification method based on sparseness filtering feature selecting |
CN107506794A (en) * | 2017-08-23 | 2017-12-22 | 西安电子工程研究所 | Ground moving object sorting algorithm based on decision tree |
CN107607925A (en) * | 2017-09-07 | 2018-01-19 | 西安电子工程研究所 | A kind of target RCS real time evaluating methods of radar application |
CN108020854A (en) * | 2017-10-26 | 2018-05-11 | 广州中南民航空管技术装备工程有限公司 | A kind of scene target battle state display method and system |
-
2018
- 2018-08-02 CN CN201810868550.3A patent/CN109117776B/en active Active
Patent Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20070024494A1 (en) * | 2005-04-11 | 2007-02-01 | Dizaji Reza M | Classification system for radar and sonar applications |
EP2420972A1 (en) * | 2010-08-19 | 2012-02-22 | BAE SYSTEMS plc | Sensor data processing |
KR20120060359A (en) * | 2010-12-02 | 2012-06-12 | 한국항공우주산업 주식회사 | Method of recognizing flight pattern of rotor craft, apparatus for recognizing flight pattern implementing the same and recording medium |
CN103885043A (en) * | 2014-03-31 | 2014-06-25 | 西安电子科技大学 | Airplane target clutter and noise stable classification method based on generalized matched filtering |
CN105242254A (en) * | 2015-10-22 | 2016-01-13 | 中国船舶重工集团公司第七二四研究所 | Air target identification method based on data quality assessment |
CN106093935A (en) * | 2016-08-26 | 2016-11-09 | 西安电子科技大学 | Pitching widebeam radar ground and low target sorting technique |
CN106483513A (en) * | 2016-09-20 | 2017-03-08 | 西安电子工程研究所 | A kind of aircraft class target micro-doppler texture characteristic extracting method |
CN107145850A (en) * | 2017-04-28 | 2017-09-08 | 西安电子科技大学 | A kind of target identification method based on sparseness filtering feature selecting |
CN107132515A (en) * | 2017-05-09 | 2017-09-05 | 哈尔滨工业大学 | A kind of point mark screening technique constrained based on multidimensional information |
CN107506794A (en) * | 2017-08-23 | 2017-12-22 | 西安电子工程研究所 | Ground moving object sorting algorithm based on decision tree |
CN107607925A (en) * | 2017-09-07 | 2018-01-19 | 西安电子工程研究所 | A kind of target RCS real time evaluating methods of radar application |
CN108020854A (en) * | 2017-10-26 | 2018-05-11 | 广州中南民航空管技术装备工程有限公司 | A kind of scene target battle state display method and system |
Non-Patent Citations (6)
Title |
---|
SIMON HAYKIN 等: "Classification of Radar Clutter in an Air Traffic Control Environment", 《IEEE》 * |
吴贝贝: "基于RCS特征提取的雷达目标识别新方法", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
宋新景 等: "低分辨雷达的目标分类识别研究", 《雷达科学与技术》 * |
张靖: "基于特征提取的目标分类研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
王旭 等: "基于点迹特征的气象杂波抑制技术", 《火控雷达技术》 * |
陈娟: "基于多特征融合的雷达目标识别", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109782247A (en) * | 2019-01-28 | 2019-05-21 | 中船重工鹏力(南京)大气海洋信息系统有限公司 | A method of utilizing track Information revision radar return |
CN109782247B (en) * | 2019-01-28 | 2020-09-22 | 中船重工鹏力(南京)大气海洋信息系统有限公司 | Method for correcting radar echo by using track information |
CN110261837A (en) * | 2019-06-27 | 2019-09-20 | 中国航空工业集团公司雷华电子技术研究所 | A kind of complex target RCS calculation method based on track information |
CN110501683A (en) * | 2019-08-19 | 2019-11-26 | 杭州电子科技大学 | A kind of extra large land Clutter Classification method based on 4 D data feature |
CN110501683B (en) * | 2019-08-19 | 2021-06-04 | 杭州电子科技大学 | Sea-land clutter classification method based on four-dimensional data characteristics |
CN111458701A (en) * | 2020-04-12 | 2020-07-28 | 西安电子工程研究所 | Meteorological track restraining method based on track characteristic iterative updating |
CN111458701B (en) * | 2020-04-12 | 2022-12-27 | 西安电子工程研究所 | Meteorological track inhibition method based on track characteristic iterative update |
CN112001342A (en) * | 2020-08-28 | 2020-11-27 | 电子科技大学 | Clutter classification method adopting VGG-16 network |
CN112001342B (en) * | 2020-08-28 | 2023-04-07 | 电子科技大学 | Clutter classification method adopting VGG-16 network |
Also Published As
Publication number | Publication date |
---|---|
CN109117776B (en) | 2021-09-07 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109117776A (en) | Aircraft and meteorological clutter classifying identification method based on track information | |
Wang et al. | Deep learning-based UAV detection in pulse-Doppler radar | |
CN109409225B (en) | Unmanned aerial vehicle classification method and device based on radar multipath signal time-frequency feature fusion | |
CN110018453A (en) | Intelligent type recognition methods based on aircraft track feature | |
CN106546975B (en) | A kind of small-sized unmanned plane based on radar data and flying bird classifying identification method | |
CA2603315A1 (en) | A classification system for radar and sonar applications | |
CN107238821A (en) | The airfield runway foreign matter detecting method and device of a kind of feature based spectrum signature | |
Chen et al. | Classification of UAV and bird target in low-altitude airspace with surveillance radar data | |
CN111142085B (en) | External radiation source radar target classification and identification method based on track feature extraction | |
CN113408328B (en) | Gesture segmentation and recognition algorithm based on millimeter wave radar | |
CN113156417B (en) | Anti-unmanned aerial vehicle detection system, method and radar equipment | |
CN107247262B (en) | A kind of airfield runway foreign matter layer detection method | |
CN111401168B (en) | Multilayer radar feature extraction and selection method for unmanned aerial vehicle | |
CN109753874A (en) | A kind of low slow small classification of radar targets method based on machine learning | |
CN113191185A (en) | Method for classifying targets of unmanned aerial vehicle by radar detection through Dense2Net | |
CN110703221A (en) | Urban low-altitude small target classification and identification system based on polarization characteristics | |
Sehgal et al. | Automatic target recognition using recurrent neural networks | |
Zhu et al. | Radar HRRP group-target recognition based on combined methods in the backgroud of sea clutter | |
CN112305509A (en) | Radar track classification and identification method based on HOG _ SVM | |
CN116415163A (en) | Unmanned aerial vehicle identification method based on radar data | |
CN110363128A (en) | Airport runway foreign matter detection method, system and medium based on biological vision | |
Sévigny et al. | Unmanned aircraft (UA) telemetry data for track modelling and classification | |
CN112327266B (en) | Clutter point trace eliminating method based on multidimensional information fusion | |
CN112418275B (en) | Millimeter wave fuse foil strip interference identification method based on range profile feature extraction | |
Zhang et al. | Research on Radar Target Recognition Algorithms |
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 |