CN111796250A - False trace point multi-dimensional hierarchical suppression method based on risk assessment - Google Patents
False trace point multi-dimensional hierarchical suppression method based on risk assessment Download PDFInfo
- Publication number
- CN111796250A CN111796250A CN202010533432.4A CN202010533432A CN111796250A CN 111796250 A CN111796250 A CN 111796250A CN 202010533432 A CN202010533432 A CN 202010533432A CN 111796250 A CN111796250 A CN 111796250A
- Authority
- CN
- China
- Prior art keywords
- trace
- point
- false
- target
- points
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 17
- 238000012502 risk assessment Methods 0.000 title claims abstract description 11
- 230000001629 suppression Effects 0.000 title claims abstract description 10
- 238000012549 training Methods 0.000 claims abstract description 19
- 238000013528 artificial neural network Methods 0.000 claims abstract description 10
- 230000007613 environmental effect Effects 0.000 claims abstract description 6
- 230000008030 elimination Effects 0.000 claims abstract description 5
- 238000003379 elimination reaction Methods 0.000 claims abstract description 5
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000012804 iterative process Methods 0.000 claims 1
- 238000001514 detection method Methods 0.000 abstract description 10
- 238000012545 processing Methods 0.000 abstract description 4
- 238000004422 calculation algorithm Methods 0.000 abstract description 3
- 239000000284 extract Substances 0.000 abstract 1
- 230000006870 function Effects 0.000 description 6
- 238000009833 condensation Methods 0.000 description 4
- 230000005494 condensation Effects 0.000 description 4
- VWDWKYIASSYTQR-UHFFFAOYSA-N sodium nitrate Chemical compound [Na+].[O-][N+]([O-])=O VWDWKYIASSYTQR-UHFFFAOYSA-N 0.000 description 3
- 238000012795 verification Methods 0.000 description 3
- 238000002592 echocardiography Methods 0.000 description 2
- 230000000977 initiatory effect Effects 0.000 description 2
- 238000001303 quality assessment method Methods 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 238000005054 agglomeration Methods 0.000 description 1
- 230000002776 aggregation Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 230000005284 excitation Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000013441 quality evaluation Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
- G01S7/417—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section involving the use of neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Computer Networks & Wireless Communication (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention relates to a multi-dimensional hierarchical suppression method for false traces based on risk assessment, and belongs to the technical field of radar data processing. Extracting multi-dimensional attribute characteristics of the agglomerated trace points, realizing the discrimination of a sample set by using an unsupervised clustering algorithm and an expert system, training the trace points and labels thereof by using an Artificial Neural Network (ANN), constructing a classifier of target trace points and environmental false alarm trace points, calculating the score of each trace point (representing the possibility generated by a target), and comparing the score of the trace point with a threshold so as to eliminate the false trace points generated by the environmental false alarm; the method maps the multidimensional characteristics of the point traces into the point trace quality, extracts the target point traces and eliminates the clutter point traces, and is simple and convenient to operate and high in elimination rate. The method solves the problem that the traditional point trace detection mode is difficult to distinguish a target point trace and a clutter point trace, improves the detection probability of small and medium-sized targets under the condition of equal false alarm rate, and reduces the possibility of track misconvergence and false tracks.
Description
Technical Field
The invention belongs to the technical field of radar data processing.
Background
The existing active radar signal detection method is based on the research that radar signals are developed under the theory of stable random white noise. And determining a detection threshold under the requirement of a false alarm rate by counting the amplitude probability density distribution of radar echoes. The typical sea clutter amplitude is lower, the EP number is larger, the saturation is large, the Doppler velocity is low, and the consistency is low; the target has larger amplitude, smaller EP number, small saturation, high Doppler velocity and high consistency. Although the sea clutter and target echoes in these dimensions have differences, the discrimination is not significant. When the traditional constant false alarm detection method adopts a unified threshold to judge, the heavy trailing detection of the sea clutter is easily output as a target signal, and the target signal is easily influenced by the clutter point trace in the track initiation and track tracking processes, so that the false initiation and the false tracking are generated. In the patent CN109613526A, spatial broadening and amplitude characteristics are adopted as multidimensional input characteristic parameters of a classifier; the patent "a three-coordinate radar trace point quality assessment method CN 108181620A" adopts a logical judgment method to perform the trace point judgment. The problems in the prior art are as follows: the radar working environment is complex, heavy tailing exists in a sea clutter area, signal variance is large, a great amount of false point tracks can be generated by adopting a constant false alarm method, so that subsequent data processing cannot effectively start and track the radar point tracks, false tracks are generated, and track error association occurs.
Disclosure of Invention
Aiming at the defects of the prior art, the invention aims to provide a multi-dimensional hierarchical suppression method for a sea-sky false target for risk assessment, which is used for performing quality assessment and false elimination on whether a point trace is generated by the target or not by utilizing the extracted point trace echo characteristics and aims to improve the target detection capability in a complex sea clutter environment.
In order to achieve the purpose, the invention adopts the following technical scheme to realize the purpose:
(1) collecting a large amount of trace point data in a training period;
(2) manually distinguishing the point traces by track tracking, unsupervised clustering and an expert system, and carrying out aggregate merging/intersection calculation to distinguish a target point trace and an environmental false alarm point trace;
(3) training the artificial neural network by using trace point data in the training set, realizing multi-dimensional characteristic risk assessment of trace points, and obtaining the similarity between the trace points and a target trace point;
(4) and according to the similarity between the trace points and the target trace points, realizing the identification and elimination of false trace points.
Specifically, the process includes a training phase and an application phase. In the training stage, extracting multidimensional attribute characteristics of space, amplitude, frequency domain, information domain and the like of the trace points after radar echo condensation, and using the normalized trace points as the input of a trace point classifier; distinguishing a sample set by using an unsupervised clustering algorithm and an expert system, labeling a large amount of trace point data, and outputting the labels as a trace point classifier; and finally, training the trace attributes and the corresponding labels thereof by using an artificial neural network ANN, learning and regularly mining the attribute characteristics of the target, and establishing a nonlinear mapping network between characteristic input and label output so as to construct a classifier of the target trace and the environmental false alarm trace.
In the application stage, target detection is carried out by adopting a traditional detection method according to the intensity of the received echo, first-stage judgment is completed, and the space, amplitude, frequency and information domain characteristics of the trace point are extracted, wherein the characteristic extraction method and the normalization method are completely the same as in the training stage; and secondly, performing second-level judgment by adopting a trained artificial neural network, inputting the extracted high-dimensional trace point characteristics, obtaining a trace point score (the probability generated by a target), realizing quality judgment, and finally, taking the trace point passing a threshold as the target generated trace point, and removing the trace point not exceeding the threshold.
The invention can adapt to complex high sea condition environment, effectively detect the target signal, and can not falsely detect the environmental echo as the target signal, thereby obviously reducing the false alarm rate. After the machine learning-based trace point quality evaluation and elimination, the false trace point number generated by the environment is remarkably reduced, the loss amount of the trace point of the target point is very small, the loss rate of the trace point of the target point in the radar visual range is less than 1 percent, the trace point generated by the environment false alarm is reduced by more than 65 percent, and the trace point quality (accuracy rate) is improved; the calculation time of the flight path is reduced by 23%, and the false flight path is reduced by more than 50%, thus proving the effectiveness of the invention.
Drawings
FIG. 1 is a multi-dimensional hierarchical false target suppression flow chart based on risk assessment.
FIG. 2 is an input and an output of a multi-dimensional hierarchical suppression of false targets based on risk assessment.
FIG. 3 is a target trace extracted after multi-dimensional hierarchical suppression of a false target.
Detailed Description
The invention is further explained below with reference to the drawings.
The used point trace input information comprises the spatial information of the point trace, such as distance, azimuth, elevation angle, threshold point trace number (EP number) during agglomeration, saturation (the ratio of the EP number to the total number of distance units in a sector), distance broadening and azimuth broadening; amplitude information such as the range bin amplitude, background amplitude of CFAR estimation, signal to noise ratio; frequency domain information such as the serial number of a main channel, the occupation ratio of condensation point traces positioned in the main channel number, the number of threshold-crossing channels, the variance of the serial number of the threshold-crossing channels and the consistency of signals in the channels; the signal processing discriminates whether or not the information is in 21 dimensions such as the region attribute (sea clutter region, ground region, noise region) and the side lobe region of the tracking target. The relevant parameters are defined as follows:
the distance refers to the distance between the current point trace and the radar
The azimuth refers to the azimuth of the current point trace in the radar coordinate system
Elevation refers to the elevation of the current point in the radar coordinate system
Amplitude refers to the average amplitude of the current trace
The background amplitude refers to the amplitude of the background where the current trace is located
The signal-to-noise ratio refers to the signal-to-noise ratio SCNR ﹦ Amp-BackAmp of the current trace point;
the threshold crossing point number refers to the threshold crossing point number forming the current trace point, and OutNum ﹦ N is recorded in the trace point condensation process;
saturation refers to the ratio of the number of threshold points forming the current trace to the trace frame
SatDeg＝N/((DisWid/DisSample)·(AziWid/AziSample))；
The distance broadening refers to the distance extension DisWid ﹦ SDis-EDis of the current trace;
the start distance SDis refers to the start distance of the current trace, the end distance EDis refers to the end distance of the current trace,
the azimuth spread refers to azimuth span AziWid ﹦ SAzi-EAzi of the current trace point;
the starting azimuth SAzi refers to the starting azimuth of the current trace point, the ending azimuth EAzi refers to the ending azimuth of the current trace point,
the region attribute refers to the type of region in which the current point trace is located, including a ground region, a sea clutter region, a cloud and rain clutter region, a noise region and an beyond-the-horizon region, the main channel number refers to a main component channel of the current point trace, it is assumed that N threshold passing points respectively belong to m output channels, and the channel numbers are respectively Chn1,…ChnmThe number of threshold-passing points of each channel is OutNums1,…OutNumsmThen the main channel number is:
MainChn1﹦Chnx，OutNumsx﹦max(OutNums1,…OutNumsm)，
the main channel proportion refers to the proportion of the threshold point of the main component channel, MainRate ﹦ OutNumsx/N，
The number of channels refers to the number of channels ChnNums ﹦ m constituting the current trace,
the channel consistency is the standard deviation ArgDeg ﹦ std (Chn) of the channel number constituting the current trace1,…Chnm)，
The side lobe mark indicates whether the current point trace is a point trace formed by a side lobe, namely the direction side lobe is 2, the distance side lobe is 1, and the distance side lobe is 0. When the trace point is located in the side lobe area of a certain trace point and meets a certain criterion, the trace point is considered as a side lobe trace point. The judgment criteria are as follows:
(1) the trace point needs to be located in a side lobe area of a certain trace point (main lobe trace point); the side lobe region is a position corresponding to the distance increase and decrease of the distance and the direction side lobe finger is at the same distance with the target and the direction increase and decrease of the direction, wherein the main lobe point trace position is taken as the center;
(2) the amplitude of the point trace and the main lobe point trace needs to satisfy the primary-secondary ratio relation;
(3) the trace point and the trace point of the main lobe need to be located in a similar output channel.
The input parameters are 21 in total.
The specific flow of the invention is shown in figure 1, and the implementation steps are as follows: step 1, forming a trace point and extracting trace point characteristics by adopting constant false alarm rate detection and trace point condensation;
step 2, associating the trace data to form a track, and selecting a track which is longer than 20 circles in association as a target track;
step 3, finding out the category serial numbers of the clustering center and the trace by a Kmeans unsupervised clustering algorithm;
step 4, sending the clustered point trace clusters into an expert system, and manually distinguishing target point traces and clutter point traces;
step 5, solving and integrating the point traces in the step 2 and the step 4 to obtain a target point trace, and solving and obtaining a clutter point trace;
step 6, dividing the trace point characteristics and the types (including label data) thereof in the step 5, randomly selecting 80% of samples as a training set, and using the rest samples as a verification set;
step 7, sending the sample set into an artificial neural network for training, and carrying out correctness verification by using a verification set;
step 8, taking the actual trace points as input, and calculating the similarity with the target trace points by using the artificial neural network trained in the step 7;
and 9, setting a trace point quantity threshold, filtering trace point data with low similarity to the target, and outputting a target trace point.
Wherein, the specific steps of the step 3 are as follows:
step 3.1: setting the number of categories to be N, wherein N is the sum of the estimated target categories and the false categories generated by the environment, for example, 80 targets represent that 80 targets are estimated, the number of categories generated by false is 20, at this time, N is selected to be 100, and when the trace quality is evaluated, a high score is given to the 80 targets and a low score is given to the 20 false targets. After N is given, N sample points are randomly selected from the point traces to serve as the class centers of the N classes; wherein X is a category attribute, the central value of the trace point attribute characteristic of the category is represented, and the trace point number is L.
Xk＝{x1，x2...xN}，i＝1，2...L，k＝1，2...N
Step 3.2: calculating Euclidean distances between all the point traces and all the categories, and judging the categories of the point traces by using the distances; where P is the characteristic dimension of the trace sample.
Tk＝arg mink{|xi-Xk|}
Step 3.3: traversing all the categories, and recalculating the centers of the categories by using all the traces of the points belonging to each category; wherein N iskRepresenting the number of trace points closest to the kth category center;
step 3.4: repeating the steps 3.2-3.3 circularly, and executing the Niter times;
step 3.5: calculating the distance between the categories by adopting the Euclidean distance, and combining the two categories of which the category distance is smaller than a threshold value into a new category; and if the distance between the mth class and the nth class is smaller than the threshold value, weighting according to the number of samples to which the distance between the mth class and the nth class belongs, calculating the center of the mth class, and deleting the centers of the mth class and the nth class.
Step 3.6: calculating the variance of each characteristic attribute in the category, splitting the category into two categories if the variance is greater than a threshold value, copying the center of the original category by the new two category centers, and adding or subtracting the standard deviation of the trace point attribute on the trace point attribute dimension (h-th attribute) with the maximum variance;
step 3.7: when the combination of classes does not occur in the step 5 and the separation of classes does not occur in the step 6, the whole unsupervised clustering process is finished, otherwise, the steps 4-6 are executed;
the specific steps of the step 7 are as follows:
step 7.1: reading the normalized input characteristic x of the data and a label v stored by an expert system;
step 7.2: setting a neural network structure and an activation function type, and randomly initializing a network weight; setting the initialization coefficients of three layers of networks as omega respectivelyiI is 1, 2, 3, and the excitation threshold is bi，i＝1，2，3；
Step 7.3: setting a network training step length h, a regularization coefficient a, a momentum coefficient M and a maximum training time Niter;
step 7.4: network forward computing to obtain the predicted result
g(x)＝x
Wherein the LeakyReLu function is:
step 7.5: calculating the difference value between the prediction result and the label truth value, namely error, defining a square cost function as the mean value of square errors of M samples, wherein a coefficient of 0.5 is a factor generated in derivation counteracting;
step 7.6: calculating backward by the network, and calculating the update coefficient of each node through a cost function and a derivative chain rule;
wherein the derivative function of the LeakyReLu function is:
step 7.7: carrying out momentum calculation on the update coefficient of the network node;
dωi＝Mdωi+Δωi
dbi＝Mdbi+Δbi，M＜1
step 7.8: forgetting the weight value of the network node slightly, and updating the weight through the result of the step 7;
ωi＝(1-α)ωi+dωi
bi＝(1-α)bi+dbi，α＜＜1
step 7.9: and 7.4-7.8 are repeated until the network error is less than the threshold value or the maximum training time Niter is reached.
Claims (3)
1. A false point trace multi-dimensional hierarchical suppression method based on risk assessment is characterized by comprising the following steps:
(1) collecting a large amount of trace point data in a training period;
(2) manually distinguishing the point traces by track tracking, unsupervised clustering and an expert system, and carrying out aggregate merging/intersection calculation to distinguish a target point trace and an environmental false alarm point trace;
(3) training the artificial neural network by using trace point data in the training set, realizing multi-dimensional characteristic risk assessment of trace points, and obtaining the similarity between the trace points and a target trace point;
(4) and according to the similarity between the trace points and the target trace points, realizing the identification and elimination of false trace points.
2. The multi-dimensional hierarchical suppression method for false traces based on risk assessment as claimed in claim 1, wherein: the unsupervised clustering is as follows: and judging the distance between the class centers in the iterative process, combining the class centers with the short distance into 1 class, and splitting the class with the large variance into 2 new classes.
3. The multi-dimensional hierarchical suppression method for false traces based on risk assessment as claimed in claim 1 or claim 2, wherein: the artificial neural network training adopts a regularization and momentum training method in the training process.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010533432.4A CN111796250A (en) | 2020-06-12 | 2020-06-12 | False trace point multi-dimensional hierarchical suppression method based on risk assessment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010533432.4A CN111796250A (en) | 2020-06-12 | 2020-06-12 | False trace point multi-dimensional hierarchical suppression method based on risk assessment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111796250A true CN111796250A (en) | 2020-10-20 |
Family
ID=72803370
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010533432.4A Pending CN111796250A (en) | 2020-06-12 | 2020-06-12 | False trace point multi-dimensional hierarchical suppression method based on risk assessment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111796250A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112327266A (en) * | 2020-10-23 | 2021-02-05 | 北京理工大学 | Clutter point trace elimination method based on multi-dimensional information fusion |
| CN112881993A (en) * | 2021-01-18 | 2021-06-01 | 零八一电子集团有限公司 | Method for automatically identifying false tracks caused by radar distribution clutter |
| CN113721211A (en) * | 2021-06-26 | 2021-11-30 | 中国船舶重工集团公司第七二三研究所 | Sparse fixed clutter recognition method based on point trace characteristic information |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109613526A (en) * | 2018-12-10 | 2019-04-12 | 航天南湖电子信息技术股份有限公司 | A kind of point mark filter method based on support vector machines |
-
2020
- 2020-06-12 CN CN202010533432.4A patent/CN111796250A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109613526A (en) * | 2018-12-10 | 2019-04-12 | 航天南湖电子信息技术股份有限公司 | A kind of point mark filter method based on support vector machines |
Non-Patent Citations (5)
| Title |
|---|
| 刘文 李宏刚 苗锋 白俊奇: "基于ReliefF 与SVM 的杂波区虚假航迹抑制方法", 指挥信息系统与技术, vol. 11, no. 2, pages 45 - 47 * |
| 刘文;李宏刚;苗锋;白俊奇;: "基于ReliefF与SVM的杂波区虚假航迹抑制方法", 指挥信息系统与技术, no. 02 * |
| 官林海;: "地面情报雷达点迹过滤技术研究", 电子技术与软件工程, no. 17 * |
| 彭威;林强;: "基于PSO-MLP的点迹真伪鉴别方法研究", 舰船电子对抗, no. 01 * |
| 林强;彭威;胡先进;: "基于改进KNN的雷达点迹真伪鉴别方法", 现代雷达, no. 04 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112327266A (en) * | 2020-10-23 | 2021-02-05 | 北京理工大学 | Clutter point trace elimination method based on multi-dimensional information fusion |
| CN112327266B (en) * | 2020-10-23 | 2024-05-03 | 北京理工大学 | Clutter point trace eliminating method based on multidimensional information fusion |
| CN112881993A (en) * | 2021-01-18 | 2021-06-01 | 零八一电子集团有限公司 | Method for automatically identifying false tracks caused by radar distribution clutter |
| CN112881993B (en) * | 2021-01-18 | 2024-02-20 | 零八一电子集团有限公司 | Method for automatically identifying false flight path caused by radar distribution clutter |
| CN113721211A (en) * | 2021-06-26 | 2021-11-30 | 中国船舶重工集团公司第七二三研究所 | Sparse fixed clutter recognition method based on point trace characteristic information |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111796250A (en) | False trace point multi-dimensional hierarchical suppression method based on risk assessment | |
| CN106599808B (en) | Hidden target extraction method based on full-waveform laser radar data | |
| CN107688170B (en) | Radar target track starting method based on random forest | |
| CN111444767B (en) | Pedestrian detection and tracking method based on laser radar | |
| CN110018453B (en) | Intelligent model identification method based on aircraft track characteristics | |
| CN111444769B (en) | Laser radar human leg detection method based on multi-scale self-adaptive random forest | |
| CN103020978A (en) | SAR (synthetic aperture radar) image change detection method combining multi-threshold segmentation with fuzzy clustering | |
| US20060072816A1 (en) | Determining statistical descriptors of a signal from a set of its samples | |
| CN111142085B (en) | External radiation source radar target classification and identification method based on track feature extraction | |
| CN112213697B (en) | Feature fusion method for radar deception jamming recognition based on Bayesian decision theory | |
| CN109597045B (en) | Static target steady recognition method based on two times of clutter suppression | |
| CN111401168A (en) | Multi-layer radar feature extraction and selection method for unmanned aerial vehicle | |
| CN113570005A (en) | Long-distance ship type identification method based on airborne photon radar | |
| CN114355298A (en) | Radar composite modulation pulse signal identification method | |
| CN115061094B (en) | Radar target recognition method based on neural network and SVM | |
| CN109283507B (en) | Radar target identification method and system based on time-frequency domain characteristics | |
| KR102361816B1 (en) | Method for detecting target and readable medium | |
| CN116660878A (en) | Active sonar target identification performance forecasting method under false target interference | |
| CN114460582B (en) | Millimeter wave radar cart identification method based on point cloud speed | |
| CN113466798B (en) | Sea surface small target detection method based on polarized signal multi-scale entropy characteristics | |
| CN115565090A (en) | Unmanned aerial vehicle unsupervised identification method and device based on radar data | |
| CN115128571A (en) | Multi-person and non-motor vehicle identification method based on millimeter wave radar | |
| CN113625266A (en) | Method, device, storage medium and equipment for detecting low-speed target by using radar | |
| CN117370750A (en) | Group target track starting method based on convolutional neural network | |
| Notkin et al. | Classification of Ground Moving Radar Targets with RBF Neural Networks. |
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 | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: 210003 No. 346, Zhongshan North Road, Jiangsu, Nanjing Applicant after: 724 Research Institute of China Shipbuilding Corp. Address before: 210003 No. 346, Zhongshan North Road, Jiangsu, Nanjing Applicant before: 724TH RESEARCH INSTITUTE OF CHINA SHIPBUILDING INDUSTRY Corp. |