CN108107410A - A kind of abnormal shape radar cascading judgement object detection method - Google Patents
A kind of abnormal shape radar cascading judgement object detection method Download PDFInfo
- Publication number
- CN108107410A CN108107410A CN201711293413.3A CN201711293413A CN108107410A CN 108107410 A CN108107410 A CN 108107410A CN 201711293413 A CN201711293413 A CN 201711293413A CN 108107410 A CN108107410 A CN 108107410A
- Authority
- CN
- China
- Prior art keywords
- radar
- mark
- detection
- range
- distance
- 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
- 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
- G01S7/021—Auxiliary means for detecting or identifying radar signals or the like, e.g. radar jamming signals
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 present invention proposes a kind of special-shaped radar cascading judgement object detection method, high-low threshold is respectively adopted, CFAR detections are carried out to every radar original video data, data extraction is apart from congealing point mark after being detected to every radar high-low threshold, according to range points mark locally letter it is miscellaneous/make an uproar than a mark amplitude is normalized with every radar noise substrate, joint-detection judgement is carried out to all radar low thresholds detection congealing point mark according to distance by radar measurement error, and multi-section radar high threshold detection congealing point mark is merged, complete the more radar joint objective detection process of original video grade.The present invention improves target detection probability using two kinds of thresholdings of height, and cascading judgement again is first extracted after range points mark for every radar, avoid more radar sampling rates, resolving power, noise floor etc. it is different caused by can not direct fusion treatment the problem of.
Description
Technical field
The invention belongs to Radar Targets'Detection technical fields.
Background technology
Increasingly complicated electromagnetic environment causes single radar target acquisition to face more and more challenges, is set up on same platform
Multi-section radar equipment task layout license under conditions of, the integration of rational time, frequency, energy etc. can be passed through
Joint objective detection is realized in scheduling, improves target acquisition ability, and has larger application potential in engineering, is mainly reflected in
Two aspects：When with the multi-section radar data on platform can be realized by way of optical fiber or network real-time Data Transmission with
Interaction, there is no data remote transmission problems needed to be considered；Second is that the multi-section radar site information on same platform
Clearly, coordinate unification, spatial synchronization and position registration are easier to realize.
Under same platform multi-section radar joint objective detection mode, more radar information Combined Treatments are to realize that detectivity carries
The important step risen, for constant false alarm rate after multi-section Radar Signal Processing (Constant False Alarm Ratio, CFAR)
Original video data before detection carries out Combined Treatment and provides possibility to promote target detection probability to greatest extent.Due to every
The parameters such as sample rate, resolving power, precision, the noise power level of portion's radar equipment may be differed, can not be directly to more thunders
It carries out merging/detection process up to original video data.In document《More radar video emerging system designs》(Xi'an electronics technology is big
Learn master thesis, 2009) method of the same target video Cluster-Fusion based on connected component labeling is proposed in, to more thunders
Space-time alignment is carried out up to video hub, the thought clustered in the identification of fusion center application mode is theoretical, by same target not
With the video hub cluster under radar, obtain subject fusion center and be sent to tracking system, but target location and quantity in practice again
It is unknown, and there are interference such as clutters, the subject fusion center that can not be confirmed the validity.In document《Low resolution based on mutual information
To extra large radar video signal registration》(high-tech communicates, 2014, Vol.24, No.8, pp:Based on two to Hai Lei in 800-806)
Up to vision signal registration spatial character and its influence factor is had studied, the target echo envelope similarities and differences between different radars are analyzed,
The method for registering based on mutual information is proposed, and is verified based on S-band and C-band measured data, but this method data
Amount of storage is big, and time-consuming for registration parameter search.
The content of the invention
The present invention is directed to many types of radar joint objective detection demand of platform, proposes a kind of special-shaped radar cascading judgement target
After multi-section radar original video data are obtained, high threshold coefficient and low threshold coefficient is respectively adopted to every thunder in detection method
Parallel C FAR detections are carried out up to data, data after every radar high-low threshold detection are obtained, according to every radar range resolution
Distance detection cohesion criterion is determined with sampling unit size, data are agglomerated into row distance after being detected to every radar high-low threshold,
Extraction apart from congealing point mark, and according to range points mark locally letter it is miscellaneous/make an uproar than the noise floor with every radar to mark amplitude into
Row normalization；All radar low thresholds detection congealing point mark is gathered into row distance according to multi-section distance by radar measurement error scope
Class carries out the cluster result for meeting cascading judgement criterion the range points mark information fusion based on normalization amplitude weighting, and right
Multi-section radar high threshold detection congealing point mark merges, and completes the more radar joint objective detection process of original video grade.
The innovative point of the present invention mainly has at 3 points, first, carrying out joint objective detection based on more radar original video level data
Processing, remains the useful information of more radar datas, improves target detection probability；Second is that using two kinds of thresholdings of height, both protected
The reliable detection of " complete weak " target is demonstrate,proved in more radar returns, also ensures the detection of " have have by force weak " target；Third, for every portion
Radar first respectively into row distance to cohesion, extracts joint-detection judgement again after range points mark, avoids more radar sampling rates, differentiates
Caused by power, noise floor etc. are different can not direct fusion treatment the problem of.
Description of the drawings
Fig. 1 is a kind of special-shaped radar cascading judgement object detection method process chart of the present invention.
Fig. 2 is the parameter that emulation is used with 2 special-shaped radar datas of platform in the embodiment of the present invention.
Fig. 3 is 1 original video data of radar that generation is emulated in the embodiment of the present invention.
Fig. 4 is 2 original video data of radar that generation is emulated in the embodiment of the present invention.
Fig. 5 is the testing result statistical chart that radar 1 individually handles emulation data in the embodiment of the present invention.
Fig. 6 is the testing result statistical chart that radar 2 individually handles emulation data in the embodiment of the present invention.
Fig. 7 is the testing result statistical chart of 2 radar Combined Treatment emulation data in the embodiment of the present invention.
Fig. 8 is to three kinds of mode handling result statistical averages in the embodiment of the present invention.
Specific embodiment
A kind of special-shaped radar cascading judgement object detection method process flow of the present invention as shown in Figure 1, with reference to flow chart and
Embodiment is specifically addressed the embodiment of the method for the present invention, and process is as follows：
Step 1：Multi-section radar data obtains.
If the multi-section radar original video data that synchronization is got are fMTP(n, i), n=1 ..., Nradar, i=
1,…,Nrange_n, wherein NradarFor radar number, N in identical platformrange_nFor the distance samples points of n-th radar.
Step 2：High-low threshold detects.
False-alarm probability according to actual needs sets high threshold detection coefficient as γ1, low threshold detection coefficient is γ2, point
It Yong not threshold coefficient γ1And γ2Parallel C FAR detections are carried out to every radar data, each range cell crosses thresholding and then retains width
Angle value, but thresholding is set to zero, obtains data f after every radar high-low threshold detectionCFAR_H(n, i) and fCFAR_L(n,i)。
Step 3：It is extracted apart from Plot coherence.
If every radar range resolution is Rres_n, unit：Rice, distance samples cell size are Δ Rn, unit：Rice, ifWhereinDownward rounding operation is represented, with Mn/MnCriterion examines every radar high-low threshold
Data f after surveyCFAR_H(n, i) and fCFAR_L(n, i) is detected into row distance sliding window to be agglomerated.Coacervation process is, when sliding window to it is a certain away from
From unit, if continuous NnRange value in a unit is all higher than 0, then distance cohesion starts, and remembers range cell serial number at this time
Idstart；Continue sliding window, work as continuous NnUnit number in a unit more than 0 is less than MnWhen, judge that distance extends and terminate, note is at this time
Range cell serial number Idend；It is N that if target, which crosses over maximum range unit number,r_TarMaxIf Idend-Idstart＞ Nr_TarMax, i.e.,
Distance extend it is excessive is determined as clutter, without apart from agglomeration process, conversely, use power centroid method calculate this mark distance forAmplitude is
Step 4：Range points mark amplitude normalization.
Calculating each range points mark, locally letter miscellaneous noise ratio isWhereinAccording to mark institute
The local clutter or noise average power that data calculate in range cell CFAR reference windows, it is equal with the noise floor of every radar
Value μnAnd standard deviation sigmanA mark amplitude is normalizedBeing multiplied by the local purpose for believing miscellaneous noise ratio is
The target normalization amplitude for keeping local letter miscellaneous noise ratio high is larger；If after the detection of n-th radar high-low threshold cohesion extraction away from
It is respectively Plot from a markn_HAnd Plotn_L。
Step 5：Low threshold detects congealing point mark cascading judgement.
If this NradarIt is up to Δ R apart from measurement system error in portion's radarmaxIf the range points of arbitrary two radars
Mark is apart from upper satisfaction | Rm,i-Rn,j|≤ΔRmax, then the two point marks are classified as a mark class Cp, it is right in this manner
The point mark Plot of all radar low threshold detection cohesionsn_L, n=1 ..., NradarIt is clustered into row distance；After completing to cluster, setting
Cascading judgement criterion Mradar/Nradar, for each class CpIf meeting criterion, i.e. NradarThere is M in portion's radarradarPortion's radar exists
Range points mark can be formed after low threshold detection, then is judged to effectively detecting；For being unsatisfactory for the class of cascading judgement criterion, delete
Except processing.
Step 6：More detections of radar result fusion outputs.
The fusion of mark information is carried out for the point mark class for meeting cascading judgement criterion in step 5, and fusion criterion is to normalize
Amplitude is weighted for weights, weight coefficient ωn,j=An,j/∑Am,l, ∑ Am,lAmplitude is normalized for all range points marks in class
The sum of, the point mark distance after Weighted Fusion is Rk=∑ ωn,jRn,j, amplitude Ak=∑ ωn,jAn,jIf the point after Weighted Fusion
Mark information is PlotL.For each point mark Plot of high threshold detection cohesionn_H, n=1 ..., Nradar, in PlotLMiddle lookup
Whether there is distance to meet | Rk-Rn,j|≤ΔRmaxPoint mark, if it is not, retaining this mark information；If so, then compare two
The normalization amplitude of person retains the larger point mark of normalization amplitude；All after traversal, the high threshold finally retained is detected
Apart from congealing point mark PlotHWith PlotLIt is exported together as cascading judgement testing result.
According to the specific embodiment of the invention, emulation generation is the same as 2 special-shaped radar original video data of platform, 2 radars
As shown in Fig. 2, certain once emulates the original video data of generation as shown in Figure 3 and Figure 4, target signal to noise ratio is disposed as parameter
7dB.Emulation carries out statistical average with 1000 groups of emulation testing results every time, and Monte Carlo are emulated 100 times, define false-alarm number
To emulate the false range points mark number of generation every time, the average false-alarm number of adjustment detection threshold control is consistent, and counts radar
1st, radar 2 and the target detection probability of joint-detection are respectively as shown in Fig. 5, Fig. 6 and Fig. 7, average detected probability and false-alarm number system
Meter is as shown in Figure 8.It can be seen from the figure that in false-alarm number under the same conditions, the target inspection of cascading judgement object detection method
Survey probability higher.
Claims (1)
1. a kind of abnormal shape radar cascading judgement object detection method, specifically includes following steps：
Step 1：Multi-section radar data obtains；
If the multi-section radar original video data that synchronization is got are fMTP(n, i), n=1 ..., Nradar, i=1 ...,
Nrange_n, wherein NradarFor radar number, N in identical platformrange_nFor the distance samples points of n-th radar；
Step 2：High-low threshold detects；
False-alarm probability according to actual needs sets high threshold detection coefficient as γ1, low threshold detection coefficient is γ2, use respectively
Threshold coefficient γ1And γ2Parallel C FAR detections are carried out to every radar data, each range cell crosses thresholding and then retains range value,
But thresholding is set to zero, obtains data f after every radar high-low threshold detectionCFAR_H(n, i) and fCFAR_L(n,i)；
Step 3：It is extracted apart from Plot coherence；
If every radar range resolution is Rres_n, unit：Rice, distance samples cell size are Δ Rn, unit：Rice, ifWhereinDownward rounding operation is represented, with Mn/MnCriterion is to every radar high-low threshold
Data f after detectionCFAR_H(n, i) and fCFAR_L(n, i) is detected into row distance sliding window to be agglomerated；Coacervation process is, when sliding window is to a certain
Range cell, if continuous NnRange value in a unit is all higher than 0, then distance cohesion starts, and remembers range cell sequence number at this time
For Idstart；Continue sliding window, work as continuous NnUnit number in a unit more than 0 is less than MnWhen, judge that distance extends and terminate, remember this
When range cell serial number Idend；It is N that if target, which crosses over maximum range unit number,r_TarMaxIf Idend-Idstart＞ Nr_TarMax,
I.e. distance extend it is excessive is determined as clutter, without apart from agglomeration process, conversely, calculating this mark distance using power centroid method
ForAmplitude is
Step 4：Range points mark amplitude normalization；
Calculating each range points mark, locally letter miscellaneous noise ratio isWhereinAccording to distance where the mark
The local clutter or noise average power that data calculate in unit CFAR reference windows, with the noise floor mean μ of every radarnWith
Standard deviation sigmanA mark amplitude is normalizedIf it is agglomerated after n-th radar high-low threshold detection
The range points mark of extraction is respectively Plotn_HAnd Plotn_L；
Step 5：Low threshold detects congealing point mark cascading judgement；
If this NradarIt is up to Δ R apart from measurement system error in portion's radarmaxIf the range points mark of arbitrary two radars exists
Apart from upper satisfaction | Rm,i-Rn,j|≤ΔRmax, then the two point marks are classified as a mark class Cp, in this manner to all
The point mark Plot of radar low threshold detection cohesionn_L, n=1 ..., NradarIt is clustered into row distance；After completing to cluster, setting joint
Decision rule Mradar/Nradar, for each class CpIf meeting criterion, i.e. NradarThere is M in portion's radarradarPortion's radar is in low door
Range points mark can be formed after limit detection, then is judged to effectively detecting；For being unsatisfactory for the class of cascading judgement criterion, make at deletion
Reason；
Step 6：More detections of radar result fusion outputs；
The fusion of mark information is carried out for the point mark class for meeting cascading judgement criterion in step 5, and fusion criterion is to normalize amplitude
It is weighted for weights, weight coefficient is For all range points marks in class normalize amplitude it
With the point mark distance after Weighted Fusion isAmplitude isIf the point mark after Weighted Fusion
Information is PlotL；For each point mark Plot of high threshold detection cohesionn_H, n=1 ..., Nradar, in PlotLMiddle lookup is
It is no to there is distance to meet | Rk-Rn,j|≤ΔRmaxPoint mark, if it is not, retaining this mark information；Both if so, then compare
Normalization amplitude, retain the larger point mark of normalization amplitude；All traversal after, by the high threshold finally retained detection away from
From congealing point mark PlotHWith PlotLIt is exported together as joint detection results.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711293413.3A CN108107410B (en) | 2017-12-08 | 2017-12-08 | Abnormal radar joint judgment target detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711293413.3A CN108107410B (en) | 2017-12-08 | 2017-12-08 | Abnormal radar joint judgment target detection method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108107410A true CN108107410A (en) | 2018-06-01 |
| CN108107410B CN108107410B (en) | 2021-05-14 |
Family
ID=62208189
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711293413.3A Active CN108107410B (en) | 2017-12-08 | 2017-12-08 | Abnormal radar joint judgment target detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108107410B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109633633A (en) * | 2018-12-27 | 2019-04-16 | 西安交通大学 | A kind of life signal Enhancement Method based on segmentation classification enhancing processing |
| CN109782247A (en) * | 2019-01-28 | 2019-05-21 | 中船重工鹏力（南京）大气海洋信息***有限公司 | A method of utilizing track Information revision radar return |
| CN114428233A (en) * | 2021-12-23 | 2022-05-03 | 西安电子科技大学 | Radar target track detection method based on multiple time and multiple resolutions |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103792522A (en) * | 2014-01-15 | 2014-05-14 | 中国人民解放军海军航空工程学院 | Multi-radar marine target robust association algorithm based on credible association pair |
| CN104749576A (en) * | 2015-04-08 | 2015-07-01 | 四川九洲空管科技有限责任公司 | Multi-radar track association and fusion method |
| US9207312B2 (en) * | 2009-09-01 | 2015-12-08 | Thales | Multi-target data processing for multi-receiver passive radars in an SFN or MFN mode |
| CN105223570A (en) * | 2015-10-22 | 2016-01-06 | 中国船舶重工集团公司第七二四研究所 | A kind of to track method for optimizing before extra large Dim targets detection |
| CN105353368A (en) * | 2015-11-09 | 2016-02-24 | 中国船舶重工集团公司第七二四研究所 | Adaptive variable structure radar sea target tracking method based on policy decision |
| CN106383344A (en) * | 2016-09-30 | 2017-02-08 | 长沙太电子科技有限公司 | Multi-station radar moving target detection method based on fusion criteria |
| CN106383339A (en) * | 2016-08-30 | 2017-02-08 | 电子科技大学 | Mirror-image object inhibition method of multi-site radar signal-level combined detection |
-
2017
- 2017-12-08 CN CN201711293413.3A patent/CN108107410B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9207312B2 (en) * | 2009-09-01 | 2015-12-08 | Thales | Multi-target data processing for multi-receiver passive radars in an SFN or MFN mode |
| CN103792522A (en) * | 2014-01-15 | 2014-05-14 | 中国人民解放军海军航空工程学院 | Multi-radar marine target robust association algorithm based on credible association pair |
| CN104749576A (en) * | 2015-04-08 | 2015-07-01 | 四川九洲空管科技有限责任公司 | Multi-radar track association and fusion method |
| CN105223570A (en) * | 2015-10-22 | 2016-01-06 | 中国船舶重工集团公司第七二四研究所 | A kind of to track method for optimizing before extra large Dim targets detection |
| CN105353368A (en) * | 2015-11-09 | 2016-02-24 | 中国船舶重工集团公司第七二四研究所 | Adaptive variable structure radar sea target tracking method based on policy decision |
| CN106383339A (en) * | 2016-08-30 | 2017-02-08 | 电子科技大学 | Mirror-image object inhibition method of multi-site radar signal-level combined detection |
| CN106383344A (en) * | 2016-09-30 | 2017-02-08 | 长沙太电子科技有限公司 | Multi-station radar moving target detection method based on fusion criteria |
Non-Patent Citations (3)
| Title |
|---|
| 杨亚波等: "雷达微弱目标检测前跟踪技术研究综述", 《雷达与对抗》 * |
| 赵德功等: "逻辑法航迹起始算法性能研究", 《雷达与对抗》 * |
| 钱骏等: "多雷达数据融合中的数据预处理", 《火控雷达技术》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109633633A (en) * | 2018-12-27 | 2019-04-16 | 西安交通大学 | A kind of life signal Enhancement Method based on segmentation classification enhancing processing |
| CN109633633B (en) * | 2018-12-27 | 2021-01-19 | 西安交通大学 | Life signal enhancement method based on segmented classification enhancement processing |
| CN109782247A (en) * | 2019-01-28 | 2019-05-21 | 中船重工鹏力（南京）大气海洋信息***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***有限公司 | Method for correcting radar echo by using track information |
| CN114428233A (en) * | 2021-12-23 | 2022-05-03 | 西安电子科技大学 | Radar target track detection method based on multiple time and multiple resolutions |
| CN114428233B (en) * | 2021-12-23 | 2024-05-17 | 西安电子科技大学 | Radar target track detection method based on multiple time and multiple resolutions |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108107410B (en) | 2021-05-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhao et al. | Bolt loosening angle detection technology using deep learning | |
| CN108107410A (en) | A kind of abnormal shape radar cascading judgement object detection method | |
| CN112147601B (en) | Sea surface small target detection method based on random forest | |
| CN109633599B (en) | Multi-target tracking method for airborne early warning radar | |
| CN104715154B (en) | Core K average Data Associations based on KMDL criterion criterions | |
| CN110443830B (en) | Precipitation cloud cluster detection method based on Gaussian mixture background model | |
| CN107783104A (en) | Tracking before a kind of more asynchronous sensor single goals detection based on particle filter | |
| CN112036239B (en) | Radar signal working mode identification method and system based on deep learning network | |
| CN111189545B (en) | High-precision wide-area intelligent infrared body temperature screening method and system | |
| Kyznarová et al. | CELLTRACK—Convective cell tracking algorithm and its use for deriving life cycle characteristics | |
| CN105372717A (en) | FOD fusion detection method and device based on radar and image signal | |
| Ningbo et al. | High-dimensional feature extraction of sea clutter and target signal for intelligent maritime monitoring network | |
| CN110889862A (en) | Combined measurement method for multi-target tracking in network transmission attack environment | |
| CN116990773A (en) | Low-speed small target detection method and device based on self-adaptive threshold and storage medium | |
| CN112327293A (en) | Sea surface target detection method based on fractal feature intelligent learning | |
| CN112213697B (en) | Feature fusion method for radar deception jamming recognition based on Bayesian decision theory | |
| Fan et al. | Weak target detection in sea clutter background using local‐multifractal spectrum with adaptive window length | |
| CN107102293A (en) | The passive co-located method of unknown clutter estimated based on sliding window integral density | |
| CN107024680A (en) | A kind of object detection method based on double deletion thresholdings | |
| CN108919255A (en) | Gao Zhongying radar weak target detection tracking based on CS-PF | |
| CN106019250B (en) | Based on angle scintillations relay type decoy discrimination method | |
| CN108507607A (en) | A kind of method for detecting weak signals based on kernel function | |
| Onumanyi et al. | A discriminant analysis-based automatic ordered statistics scheme for radar systems | |
| CN109766737B (en) | Image processing-based indoor human body posture preliminary classification method | |
| Chen | A corner matching algorithm based on Harris operator |
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 |