CN104730509B - A kind of radar detecting method of knowledge based complement replacement inspection - Google Patents
A kind of radar detecting method of knowledge based complement replacement inspection Download PDFInfo
- Publication number
- CN104730509B CN104730509B CN201510162189.9A CN201510162189A CN104730509B CN 104730509 B CN104730509 B CN 104730509B CN 201510162189 A CN201510162189 A CN 201510162189A CN 104730509 B CN104730509 B CN 104730509B
- Authority
- CN
- China
- Prior art keywords
- unit
- reference unit
- available strategy
- matrix
- thresholding
- 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.)
- Expired - Fee Related
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/28—Details of pulse systems
- G01S7/285—Receivers
- G01S7/292—Extracting wanted echo-signals
- G01S7/2923—Extracting wanted echo-signals based on data belonging to a number of consecutive radar periods
- G01S7/2928—Random or non-synchronous interference pulse cancellers
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention discloses a kind of radar detecting method of knowledge based complement replacement inspection, step is：(1) false-alarm probability and the basic parameter such as reference unit and pulse number according to setting, calculates sequence numbers of the thresholding when all displacement statistics press descending arrangement, and then determines all of available strategy and thus composition available strategy matrix；(2) effective reference unit is selected according to the GIS information of unit to be detected and reference unit, and valid data matrix is constituted by the sampled value of effective reference unit and unit to be detected；(3) detection statistic is calculated according to the data of unit to be detected, replace statistic according to available strategy matrix calculus and further determine thresholding, detection statistic and thresholding are compared so as to complete judgement.The present invention effectively improves performance of the algorithm under complex clutter environment；The present invention greatly expands the scope of application；The present invention significantly reduces amount of calculation, improves target detection speed.
Description
Technical field
The invention belongs to Radar Targets'Detection field, is related to a kind of detections of radar side of knowledge based complement replacement inspection
Method.
Background technology
CFAR (CFAR) detection refers to that a class can keep constant false-alarm general under conditions of noise level is continually changing
The target detection technique of rate.Common CFAR detection algorithms can be divided into two classes：Parameteric CFAR and nonparametric CFAR.Parameteric CFAR
Algorithm usually assumes that the distribution pattern of clutter is known, it is only necessary to estimate some unknown parameters, and design thresholding accordingly to make mesh
Mark detection has CFAR characteristic under the assumptions.Nonparametric CFAR algorithms require no knowledge about the distribution pattern of clutter, and which is
The inspection policies with constant false-alarm probability under the weak hypothesis with regard to background noise or clutter statistical characteristicses.In the non-equal of complexity
In even clutter environment, these traditional CFAR detection techniques are generally difficult to obtain preferable performance.So CFAR researchs in recent years
An important directions be exactly, by means of knowledge assistance, fully to merge various prior informations (such as target movable information, meteorology
Information, road traffic map and geography information figure etc.), design the CFAR algorithms of self adaptation complex clutter environment.A.De Maio et al.
Using various possible prior informations, the Parameteric CFAR detection algorithm of several knowledge based auxiliary is devised.It is non-homogeneous processing
During clutter environment, the adaptive detection algorithm of these knowledge based auxiliary is because increased data screening process, and shows substantially
Better than the performance of traditional CFAR detections.
Current knowledge assistance CFAR algorithms mainly merge prior information shape on the basis of the detection of common Parameteric CFAR
Into, thus need also exist for assume clutter distribution pattern known to, it means which is only applicable to a group distribution function has
Specify the clutter type of analytic expression.In practical application, this will be a serious unfavorable factor, cause knowledge assistance Parameteric CFAR
The scope of application of algorithm is limited significantly, because in most cases the distribution pattern of clutter is unknown or cannot use concrete function table
Show.
The content of the invention
The technical problem to be solved in the present invention is：For traditional CFAR algorithms penalty in non-homogeneous clutter environment,
The problem of new knowledge assistance Parameteric CFAR algorithm scope of application critical constraints, proposes a kind of knowledge based complement replacement inspection
Radar detecting method, the method are realized under complex clutter environment by the permutation test cascade by data selector with nonparametric
Excellent detection performance, while with the wide scope of application.
The technical solution adopted for the present invention to solve the technical problems is：A kind of radar of knowledge based complement replacement inspection
Detection method, the method realize that step is as follows：
Step (1), the false-alarm probability according to setting and the basic parameter such as reference unit and pulse number, calculate thresholding and exist
All displacement statistics press sequence number during descending arrangement, and then determine all of available strategy and thus constitute available strategy
Matrix；
Step (2), effective reference unit is selected according to the GIS information of unit to be detected and reference unit, and by
Effectively the sampled value of reference unit and unit to be detected constitutes valid data matrix；
Step (3), according to the data of unit to be detected calculate detection statistic, according to available strategy matrix calculus displacement unite
Metering simultaneously further determines thresholding, and detection statistic and thresholding are compared so as to complete judgement.
Further, using the structure that data selector cascades permutation test, the method realizes that data selector is to reference
After the data of unit are screened, only retain those reference datas similar to element characteristics to be detected, then recycle these
Reference data carries out object detection process by permutation test.
Further, the base such as false-alarm probability and reference unit and pulse number according to setting in the step (1)
This parameter, calculates sequence number of the thresholding when all displacement statistics press descending arrangement, and then determines all of available strategy
And available strategy matrix is thus constituted, the step is preprocessing process, only needs to perform when target detection is carried out to multiple units
Once and acquired results being stored, being carried out once during without detecting to each unit.
Further, the method for discrimination of the concept and " available strategy " with regard to " available strategy " is this method proposed, and
Thresholding is determined using the corresponding displacement statistic of available strategy.
Advantage is the present invention compared with prior art：
1) by introduce priori GIS information reference unit is screened, select have to unit to be detected it is similar
The reference data of characteristic, effectively improves performance of the algorithm under complex clutter environment；
2) the target detection stage using nonparametric permutation test process, make the present invention algorithm can apply to it is various miscellaneous
Ripple distribution pattern it is unknown or cannot be with concrete function representation in the case of, greatly expand the scope of application；
3) in the permutation test stage, a kind of brand-new implementation method is employed, amount of calculation is significantly reduced, is improve mesh
Mark detection speed.
Description of the drawings
Fig. 1 is observation data model conventional in Radar Targets'Detection；
Fig. 2 is a kind of particular flow sheet of the radar detecting method of knowledge based complement replacement inspection of the invention.
Specific embodiment
Below in conjunction with the accompanying drawings and specific embodiment is discussed in detail the present invention.
GIS information (GIS is GIS-Geographic Information System) is applied in Radar Targets'Detection in the present invention, it is proposed that data are selected
Select the detection method that device cascades permutation test, the detection performance being effectively improved in complex clutter environment；Meanwhile, devise one
Brand-new permutation test concrete methods of realizing is planted, the amount of calculation of permutation test is greatly reduced.
A kind of radar detecting method of knowledge based complement replacement inspection of the present invention, which is embodied as flow process such as Fig. 2 institutes
Show, specifically comprising following 3 steps：
1st, according to basic parameters such as the false-alarm probabilities and reference unit and pulse number for setting, calculate thresholding and put all
Statistic is changed by sequence number during descending arrangement, and then is determined all of available strategy and is thus constituted available strategy matrix.
Permutation test is a kind of typical nonparametric CFAR detection algorithms so as to keep constant false alarm rate only to need the connection of clutter
Close distribution function and meet " permutation invariance ", and the concrete distribution pattern of clutter need not be known.Therefore, in relevant clutter type
In the case that prior information is less, permutation test has a wider suitability, but permutation test to there is also amount of calculation excessive etc.
Problem.
Consider pulse radar observation model as shown in Figure 1：Whole observation area includes unit to be detected and its both sides pair
Claim M reference unit of distribution, and carry out altogether n times independent scan.Adopt i ＆ lt scanning is carried out to all observing units
Sample is denoted as：
xi=(xi,1,xi,2,…,xi,M/2,xi,0,xi,M/2+1,…,xi,M)T (1)
Wherein, xi0For the sampling of unit to be detected, xi1,xi2,…,xiMFor the sampling of reference unit, i=1,2 ..., N.In
It is that all observation data can be denoted as the matrix of (M+1) × N-dimensionalFor simplicity, by the row of matrix X
Vector is designated as respectively, r0=(x1,0,x2,0,…,xN,0) it is the sampling for treating the scanning of detector unit n times；rj=(x1,j,x2,j,…,
xN,j) it is sampling to the scanning of j-th reference unit n times, j=1,2 ..., M.
Define H0It is assumed to be at unit to be detected and there is no target, H1It is assumed to be at unit to be detected and there is target.In H0It is false
Set, vector xi, each component of i=1,2 ..., N is independent same distribution (IID)；And in H1Under hypothesis, only xi1,xi2,…,
xiMIt is IID.By H0And H1Lower stochastic variable xi0Probability density function be denoted as f respectively0i(xi0) and f1i(xi0), i=1,
2,…,N.Because two kinds are assumed at lower reference unit all the time without target, xijProbability density function always be f0i(xij), i=
1,2 ..., N and j=1,2 ..., M.Thus, the probability density function of matrix X is：
Formula (3) and formula (2) are made to compare and take the logarithm, can obtain log-likelihood ratio (i.e. detection statistic) is：
Wherein,
From each column x of matrix XiIn appoint take a sampled valueAnd with the x in its alternate form (4)i0(this is corresponding in X
WillWith xi0Displacement), it is obtained：
Above formula referred to as replaces statistic.Wherein, k=(k1,k2,…,kN) and ki∈ { 0,1 ..., M }, i=1,2 ..., N,Obviously, formula (4) be formula (5) k=(0,0 ..., special case when 0).
For observing matrix X=(x1,x2,…,xN), it is different according to the value of k, can obtain (M+1) altogetherNIndividual difference
Tk() value.By these Tk() is arranged according to descending order, and takes Q as thresholding(its
In, Q is by false-alarm probability P for settingfaThe value of decision).Then, permutation test can be expressed as：
In H0Under hypothesis, vector xi, each component of i=1,2 ..., N meets IID characteristics, then all (M+1)NIndividual Tk(·)
It is that equiprobability occurs, then false-alarm probability is：
It can be seen that, false-alarm probability PfaOnly it is the function of parameter M, N and Q, and it is unrelated with clutter distribution function, and this exactly replaces inspection
Test the reason for constant false alarm rate being kept by need not knowing clutter distribution pattern.
Can be directly obtained according to formula (7)：
Q=Pfa×(M+1)N (8)
This is sequence numbers of the thresholding T in all displacement statistics of descending arrangement in traditional permutation test.And
In for a kind of radar detecting method of knowledge based complement replacement inspection of the invention, the data of original reference unit need elder generation Jing
Data selector is used further to permutation test after being screened.Assume that the reference unit number that data selector retains Jing after screening is
Ms, then in follow-up permutation test, sequence numbers of the thresholding T in all displacement statistics should be：
Qs=Pfa×(Ms+1)N (9)
In the present invention, (a M can be exported after data selector is screened to sampled datas+ 1) × N-dimensional it is effective
Data matrix Xs, the sampled value of the first behavior unit to be detected of the matrix, other behaviors selected reference unit sampled value.
By matrix XsEach element press ai(xij)=ln (f1i(xij)/f0i(xij)) carry out process and obtain matrix As, then to AsEach row unit
Element carries out descending arrangement, you can obtain matrixObviously,And (Ms+ 1) which specifically might as well be denoted as by × N：
From matrixEach row in choose one-component respectively and obtain one group(wherein, mi∈{1,
2,…,Ms+ 1 }, i=1,2 ..., N), thus can try to achieve a displacement statisticThe process is corresponding to arrow
AmountAgain due to matrixEach column element be descending arrangement, if so m=(m1,m2,…,mN)
With n=(n1,n2,…,nN) meet ni≤miTo i=1,2 ..., N sets up, then necessarily haveThen,
For vector m, can at least findIt is individual to be not less thanStatistic.IfThenNecessarily
Not in front QsThe row of individual maximum statistic, at this moment m is called " invalidation policy "；Conversely,Front Q may be belonged tosIndividual maximum
One of statistic, m are called " available strategy ".Using all of available strategy as a line, you can constitute available strategy matrix.
As the corresponding statistic of any invalidation policy is not in front QsThe row of individual maximum statistic, so thresholding T is available strategy
Q in the corresponding displacement statistic of matrixsThat big statistic.
Analyze more than comprehensive, determine that available strategy matrix can specifically be divided into two steps：
1. calculating formula Qs=Pfa×(Ms+1)N。
2. to each policy vector m=(m1,m2,…,mN), mi∈{1,2,…,Ms+ 1 }, i=1 ..., N, ifVector m is judged to into available strategy then.Using all available strategies as a line, available strategy matrix is constituted.
It should be noted that QsAnd available strategy matrix only with parameter Pfa、MsIt is relevant with N, and with radar observation unit
Concrete sampled value is unrelated.When target detection is carried out to multiple observing units, due to parameter Pfa、MsImmobilize with N, so can
Using the 1st step of the invention as preprocessing process, result is only performed once and is stored before detection starts, without to each unit
It is carried out during detection.So process and can substantially reduce amount of calculation.
2nd, effective reference unit is selected according to the GIS information of unit to be detected and reference unit, and by effectively joining
The sampled value for examining unit and unit to be detected constitutes valid data matrix.
In the present invention, data selector needs to screen reference unit using the GIS information of priori.Here it is used
The GIS information for arriving, refers to the grid that observation area is divided into fixed size, is then represented corresponding to each grid with numeral
The matrix of geographical feature (such as land, ocean, woods etc.).In general, the region with identical geographical feature, its noise performance
Also it is close, this exactly in the present invention data selector basis of uniform clutter unit is filtered out using GIS.
Here using the fixed number selector for being easier to realize, also known as FKA data selectors, the selector can be from
Fixed value M is selected in M initial reference unitsIndividual reference unit.For simplicity, consider first GIS grids and distance by radar-
The one-to-one situation of localizer unit.The GIS information of unit to be detected (CUT) and reference unit is denoted as into Y respectively0And Yj, j=
1 ..., M, and defined function：
IfThen effective reference unit is selected to meet D (j)=1 and apart from CUT nearest MsIt is individual
Reference unit；IfThen effective reference unit is selected as all reference units and D for meeting D (j)=1
It is nearest apart from CUT in (j)=0Individual reference unit.Using the data of unit to be detected as the first row, often
The data of individual effective reference unit are also individually embarked on journey, and constitute (a Ms+ 1) matrix of × N-dimensional is valid data matrix Xs。
In some cases, the resolution of GIS may be more many than the high resolution of radar, such a distance by radar-orientation
Multiple GIS units will be included in unit.Now, data selector will become somewhat more complex.It is assumed that the geographical letter of a total of L kinds
Breath type, numbering are respectively 1,2 ..., L.To each radar observation unit, statistics wherein numbering is the GIS of l ∈ { 1,2 ..., L }
Unit number, you can obtain a L n dimensional vector n v, as the geography information characteristic vector of the radar cell.Then, jth ∈
{ 1,2 ..., M } individual reference unit can be represented with following formula with the geography information similarity of unit to be detected：
Wherein, vjAnd v0The geography information characteristic vector of respectively j-th reference unit and unit to be detected,<·>Represent
Inner product operation, | | | | represent vector modulus.In this case, FKA data selectors need to only select the M for making SIM maximumsIndividual ginseng
Examine unit.
In sum, determine that valid data matrix can specifically be divided into following three step：
1. the geography information similarity between each reference unit and unit to be detected is determined according to formula (10) or formula (11)；
2. M is selected from all reference unitssIt is individual with unit geography information similarity highest unit conduct to be detected
Effective reference unit (if similarity is identical, the nearer reference unit of chosen distance unit to be detected)；
3. using the data of unit to be detected as the first row, the data of each effective reference unit are also individually embarked on journey, and constitute
One (Ms+ 1) matrix of × N-dimensional is valid data matrix Xs。
3rd, detection statistic is calculated according to the data of unit to be detected, replace statistic simultaneously according to available strategy matrix calculus
And then determine thresholding, detection statistic and thresholding are compared so as to complete judgement.
Valid data matrix X is obtained by data selectorsAfterwards, it is necessary to carry out target detection with permutation test.
When carrying out permutation test using direct computing method, need to calculate all of (Ms+1)NIndividual displacement statistic Tk(), then therefrom select
Select QsIndividual maximum is used as thresholding.For slightly bigger MsAnd N, this will be very huge amount of calculation.In view of this
Point, devises a kind of brand-new permutation test implementation method in the present invention, and the method only needs to calculate all available strategies corresponding
Displacement statistic, thus also referred to as " available strategy " method.
It is assumed that what FKA data selectors selected is theM altogethersIndividual reference unit, then effectively observe data
Matrix can be expressed as：
Wherein, the sampling of detector unit n times scanning is treated in the first behavior, and remaining row is that effective reference unit n times are scanned
Sampling.Q has been tried to achieve in the 1st step of the inventionsUnder the precondition of available strategy matrix, the concrete mistake of the implementation method
Journey is as follows：
1. according to matrix XsCalculate new matrix As
Wherein, aij=ai(xij)=ln (f1i(xij)/f0i(xij)), i=1,2 ..., N andThis
Outward, under Gaussian Clutter background and fixed target conditions, quadratic detection can be reduced toOr linear detection aij=
|xij|。
2. to matrix AsThe first row element summation, obtain detection statistic
3. by matrix AsEach column element arrange in descending order, obtain new matrix
Wherein,
4. according to matrixThe corresponding displacement statistic of each available strategy in available strategy matrix is calculated, then by which
Descending is arranged, and selects QsIt is individual as thresholding T.
5. compare detection statistic T (X) and thresholding T, make judgement.
As can be seen that the whether available strategy method that proposes in direct computing method or the present invention, main amount of calculation is all
It is displacement statistic TkThe calculating of ().Therefore, it can the displacement statistic number of required calculating as amount of calculation size
Mark.For available strategy method, need to calculate the corresponding statistic of all available strategies, its number is：
Wherein, mi1,2 are taken equiprobably ..., Ms+ 1, i=1,2 ..., N,RepresentIt is general
Rate.And direct computing method needs to calculate all statistics, its number is：
Ndir=(Ms+1)N (13)
Formula (12) and formula (13) seek ratio, can obtain：
Above formula can approximately regard the ratio of the amount of calculation of both implementation methods as.Its value is less, shows available strategy method efficiency
It is higher.Set up as γ≤1 is permanent again, so available strategy method necessarily will not be poorer than direct computing method.In fact, in QsIt is less
In the case of, the amount of calculation of available strategy method will be much smaller than direct computing method.
The present invention is verified below by the method for emulation.The data that adopt of experiment for：From Canadian Mcmaster
The front N intercepted in IPIX radars actual measurement file 13r=102 range cells, 220 ° to 304 ° common NaThe number of=1834 orientations
According to.In object detection field, IPIX radar datas disclosed in Mcmaster universities are the conventional datas for evaluation algorithms performance
One of.The platform configuration that adopts of experiment for：The CPU of Intel Core i7-3770,4GB internal memories.The parameter setting that experiment is adopted
For：False-alarm probability Pfa=10-6, reference window length M=12, multiple scanning times N=7, reference unit selected by data selector
Number Ms=9.The detections of radar side of knowledge based complement replacement inspection in traditional permutation test method and the present invention is respectively adopted
Method carries out false-alarm detection to observation area, and the former false-alarm number is 638, and the false-alarm number of the latter is 99.Then, in observation
In each orientation of data, a range gate insertion point target is randomly selected, and mesh is carried out with both detection methods respectively
Mark detection.As a result show, in the present invention, the radar detecting method of knowledge based complement replacement inspection is than traditional permutation test
Method has the advantage of about 3dB.Finally, direct computing method is respectively adopted and available strategy method is aided in as knowledge based of the present invention
100 observing units are detected by the concrete methods of realizing of permutation test part in the radar detecting method of permutation test.Directly
The time connect needed for calculating method is 2018s, and the time needed for available strategy method is only 0.26s.
The techniques well known being related in the present invention is not elaborated.
Claims (3)
1. the radar detecting method that a kind of knowledge based complement replacement is checked, it is characterised in that realize that step is as follows：
Step (1), the false-alarm probability according to setting and the basic parameter such as reference unit and pulse number, calculate thresholding all
Displacement statistic presses sequence number during descending arrangement, and then determines all of available strategy and thus constitute available strategy square
Battle array；Determine that available strategy matrix can specifically be divided into two steps：
1. calculating formula Qs=Pfa×(Ms+1)N；PfaFor false-alarm probability, QsIn follow-up permutation test, thresholding T is in all displacement statistics
Sequence number in amount, MsFor the reference unit number that data selector retains Jing after screening；
2. to each policy vector m=(m1,m2,…,mN), mi∈{1,2,…,Ms+ 1 }, i=1 ..., N, ifThen will
Vector m is judged to available strategy, using all available strategies as a line, constitutes available strategy matrix；
Step (2), effective reference unit is selected according to the GIS information of unit to be detected and reference unit, and by effective
The sampled value of reference unit and unit to be detected constitutes valid data matrix；
Step (3), according to the data of unit to be detected calculate detection statistic, according to available strategy matrix calculus replace statistic
And and then determine thresholding, detection statistic and thresholding are compared so as to complete judgement.
2. the radar detecting method that knowledge based complement replacement according to claim 1 is checked, it is characterised in that：The method
The structure realization of permutation test is cascaded using data selector, after data selector is screened to the data of reference unit, only
Retain those reference datas similar to element characteristics to be detected, then recycle these reference datas to carry out by permutation test
Object detection process.
3. the radar detecting method that knowledge based complement replacement according to claim 1 is checked, it is characterised in that：The step
Suddenly the basic parameter such as false-alarm probability and reference unit and pulse number according to setting in (1), calculates thresholding and puts all
Statistic is changed by sequence number during descending arrangement, and then is determined all of available strategy and is thus constituted available strategy matrix,
The step is preprocessing process, only needs to perform once and store acquired results when target detection is carried out to multiple units,
It is carried out once during without detecting to each unit.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510162189.9A CN104730509B (en) | 2015-04-08 | 2015-04-08 | A kind of radar detecting method of knowledge based complement replacement inspection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510162189.9A CN104730509B (en) | 2015-04-08 | 2015-04-08 | A kind of radar detecting method of knowledge based complement replacement inspection |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104730509A CN104730509A (en) | 2015-06-24 |
| CN104730509B true CN104730509B (en) | 2017-03-29 |
Family
ID=53454588
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510162189.9A Expired - Fee Related CN104730509B (en) | 2015-04-08 | 2015-04-08 | A kind of radar detecting method of knowledge based complement replacement inspection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104730509B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105182312B (en) * | 2015-09-29 | 2017-09-12 | 大连楼兰科技股份有限公司 | The CFAR detection method of adaptive environment change |
| CN106291495B (en) * | 2016-07-26 | 2019-03-19 | 中国船舶重工集团公司第七二四研究所 | It is a kind of based on arteries and veins group interlock layout improve target signal to noise ratio scan method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5416488A (en) * | 1993-12-27 | 1995-05-16 | Motorola, Inc. | Radar return signal processing method |
| CN101975940A (en) * | 2010-09-27 | 2011-02-16 | 北京理工大学 | Segmentation combination-based adaptive constant false alarm rate target detection method for SAR image |
| CN104316914A (en) * | 2014-11-03 | 2015-01-28 | 西安电子科技大学 | Radar target self-adaptation detection method depending on shape parameters |
-
2015
- 2015-04-08 CN CN201510162189.9A patent/CN104730509B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5416488A (en) * | 1993-12-27 | 1995-05-16 | Motorola, Inc. | Radar return signal processing method |
| CN101975940A (en) * | 2010-09-27 | 2011-02-16 | 北京理工大学 | Segmentation combination-based adaptive constant false alarm rate target detection method for SAR image |
| CN104316914A (en) * | 2014-11-03 | 2015-01-28 | 西安电子科技大学 | Radar target self-adaptation detection method depending on shape parameters |
Non-Patent Citations (2)
| Title |
|---|
| An Adaptive CFAR Detector for Range-Spread Target in Various Environments;Ran Shiling et al.;《Proceedings of 2012 IEEE 11th International Conference on Signal Processing(ICSP 2012)》;20121021;第1980-1983页 * |
| 一种基于知识辅助的CFAR检测器;李青华等;《雷达科学与技术》;20120215;第10卷(第1期);第88-93页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104730509A (en) | 2015-06-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Goyal et al. | Identification of homogeneous rainfall regimes in Northeast Region of India using fuzzy cluster analysis | |
| CN107103332B (en) | A kind of Method Using Relevance Vector Machine sorting technique towards large-scale dataset | |
| CN103489005B (en) | A kind of Classification of High Resolution Satellite Images method based on multiple Classifiers Combination | |
| CN107633226B (en) | Human body motion tracking feature processing method | |
| CN108647732B (en) | Pathological image classification method and device based on deep neural network | |
| CN106874688A (en) | Intelligent lead compound based on convolutional neural networks finds method | |
| CN105677791B (en) | For analyzing the method and system of the operation data of wind power generating set | |
| CN105549009B (en) | A kind of SAR image CFAR object detection methods based on super-pixel | |
| CN108694390A (en) | A kind of cuckoo search improves the modulated signal sorting technique of grey wolf Support Vector Machines Optimized | |
| CN105225227B (en) | The method and system of remote sensing image change detection | |
| CN106093946A (en) | A kind of target condensing method being applicable to scene surveillance radar and device | |
| CN106295498A (en) | Remote sensing image target area detection apparatus and method | |
| CN101908213A (en) | SAR image change detection method based on quantum-inspired immune clone | |
| US20220028068A1 (en) | Multi-scale tumor cell detection and classification | |
| CN111815640B (en) | Memristor-based RBF neural network medical image segmentation algorithm | |
| CN104730509B (en) | A kind of radar detecting method of knowledge based complement replacement inspection | |
| CN107766792B (en) | Remote sensing image ship target identification method | |
| CN104537384B (en) | A kind of SAR target discrimination methods of combination likelihood ratio judgement | |
| CN113096080B (en) | Image analysis method and system | |
| CN110472607A (en) | A kind of ship tracking method and system | |
| CN110166422A (en) | Domain name Activity recognition method, apparatus, readable storage medium storing program for executing and computer equipment | |
| CN111104339A (en) | Software interface element detection method and system based on multi-granularity learning, computer equipment and storage medium | |
| CN111949530B (en) | Test result prediction method and device, computer equipment and storage medium | |
| CN112859034B (en) | Natural environment radar echo amplitude model classification method and device | |
| Zhang et al. | Non‐parametric detector in non‐homogeneous clutter environments with knowledge‐aided permutation test |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170329 Termination date: 20180408 |