CN116699617A - Floating target identification method by utilizing centroid time sequence information - Google Patents
Floating target identification method by utilizing centroid time sequence information Download PDFInfo
- Publication number
- CN116699617A CN116699617A CN202310576670.7A CN202310576670A CN116699617A CN 116699617 A CN116699617 A CN 116699617A CN 202310576670 A CN202310576670 A CN 202310576670A CN 116699617 A CN116699617 A CN 116699617A
- Authority
- CN
- China
- Prior art keywords
- target
- centroid
- doppler spectrum
- sequence
- standard1
- 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 39
- 238000001228 spectrum Methods 0.000 claims abstract description 53
- 230000002159 abnormal effect Effects 0.000 claims abstract description 17
- 238000012545 processing Methods 0.000 claims abstract description 5
- 238000012549 training Methods 0.000 claims description 15
- 239000013598 vector Substances 0.000 claims description 12
- 238000012937 correction Methods 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 6
- 230000006870 function Effects 0.000 claims description 6
- 238000001514 detection method Methods 0.000 claims description 5
- 238000010606 normalization Methods 0.000 claims description 4
- 230000001427 coherent effect Effects 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 3
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 claims description 2
- 238000002592 echocardiography Methods 0.000 claims description 2
- 238000010183 spectrum analysis Methods 0.000 claims description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention relates to the technical field of radar signal processing, in particular to a method for identifying a sea surface floating target and a ship. The Doppler spectrum centroid time sequence evolution rule information of the floating target is utilized, the target echo information utilization rate is improved, and the problem that the target cannot be finely identified by only depending on the one-dimensional range profile characteristics of the target in the past is solved; the invention provides a double-threshold method for eliminating abnormal values of a centroid sequence, which is suitable for the common abnormal value eliminating situation; on the second-level observation time, the accuracy of the method provided by the invention for identifying the targets with larger mass difference, such as ship targets, common floating targets and the like, can reach more than 90%.
Description
Technical Field
The invention relates to the technical field of radar signal processing, in particular to a method for identifying a sea surface floating target.
Background
On the premise that the sea surface target can be effectively detected by the sea detection radar, the sea surface target is accurately identified and classified to determine the specific category, the application and the threat level, and the method has important significance in the civil field and the military field. Common sea surface floating targets mainly comprise small fishing vessels, buoys, floating ices and the like, and the small fishing vessels, buoys, floating ices and the like are generally different in structure and physical size from ship targets such as large pleasure vessels, cargo rolling vessels and the like. The high-resolution radar can obtain the one-dimensional range profile characteristics of the target, reflects the distribution of strong scattering points along the distance, is an important target characteristic, and is a main basis for identifying a floating target and a large ship. However, the one-dimensional range profile features are closely related to the observation angle, and particularly when the radar beam irradiates along the side of the ship, the ship target and the floating target have similar one-dimensional range profile features, which seriously affect the recognition performance. In addition, for low resolution system radars, it is particularly difficult to identify floating targets without relying on ancillary information, since one-dimensional range profile features are not available.
The characteristic of difference between the ship target and the floating target is deeply excavated from the physical mechanism of the ship target and the floating target, and the method is a fundamental approach for solving the problems faced when the one-dimensional range profile method is used for identifying the floating target. Obviously, the general density and mass of the ship target are larger, and the general density and mass of the floating target are smaller, so that the ship target is always in a stable state and is rarely fluctuated with sea waves in a severe manner under the general sea condition, and the floating target is always fluctuated up and down with the sea waves.
Therefore, the ship target and the floating target have different fluctuation characteristics, so the invention aims to distinguish the ship target and the floating target by utilizing Doppler information of target radar echo.
Disclosure of Invention
The invention aims to provide a floating target identification method utilizing Doppler spectrum centroid time sequence information, in particular to a sea surface floating target and a ship target which cannot be distinguished by a low-resolution radar and cannot acquire a one-dimensional range profile or have similar physical dimensions.
The method for identifying the floating target by utilizing the centroid time sequence information is characterized by comprising the following steps of:
step 1, extraction of Doppler spectrum centroid sequence
Sea detection radar detects sea surface targets, and a coherent pulse string x (N) with the length of N is received on a certain target distance unit, wherein n=1, 2, … and N. Cut x (N), n=1, 2, …, N into non-overlapping short vectors u of length I i The following is shown:
[x(1),x(1),…,x(N)] T ＝[u 1 ,u 2 ,…,u N/I ] T (1)
respectively calculating each short vector u i And extracting doppler spectrum centroid features. The invention uses the power spectrum as Doppler spectrum, and adopts a non-parameter method in modern signal spectrum analysis to estimate average and short-time Doppler spectrum. To effectively reduce side lobes, a windowing process is used in the estimation.
The Doppler spectrum estimation method comprises the following steps: assuming that the length of a single data segment is L, for the phi (phi=1, 2, …, phi) data segment, the short-time doppler spectrum estimate is expressed as:
wherein , and />Represents a short-time magnitude spectrum and a short-time power spectrum, respectively, ω (n) represents a time-domain window function, T represents a pulse repetition period (Pulse repetition interval, PRI), c (φ) (n), n=1, 2, …, L represents the radar echo sequence of the phi-th data segment, delta is the window function power, expressed as:
and taking an average value of the short-time Doppler spectrum along the frequency dimension to obtain an average Doppler spectrum estimated value, wherein the average Doppler spectrum estimated value is expressed as:
doppler centroid f C Describing the Doppler shift degree of the target, the estimation method is as follows:
where Q (Φ) represents the power level of the short-term doppler spectrum for the Φ data segment, namely:
the Doppler spectrum centroid sequence of a certain sea surface target is finally obtained and recorded as X 1 ,X 2 ,…,X N/I 。
Step 2, outlier correction of centroid sequence
The Doppler spectrum centroid sequence extracted in the step 1 always has abnormal values due to the random interference of the same-frequency interference, noise, unknown source electromagnetic waves and the like. Doppler spectrum centroid sequence X of sea surface target by adopting double-threshold method 1 ,X 2 ,…,X N/I The process flow is as follows:
1) Initializing: according to the approximate change trend of the centroid sequence, two decision threshold standards 1 and 2 are set. When the dimension of the centroid is the normalized frequency, the value of standard1 ranges from 0.03 to 0.15, and the value of standard2 ranges from 0.05 to 0.3, for example, standard 1=0.03 and standard 2=0.05.
2)X 1 Is characterized by comprising the following steps: judging |X 1 -mean([X 2 ,X 3 ]) Whether the absolute value is smaller than standard1, wherein mean (·) represents mean operation, and the absolute value is represented. If |X 1 -mean([X 2 ,X 3 ])|<Standard1, X 1 Normal value, otherwise X 1 Let X be an outlier 1 ＝mean([X 2 ,X 3 ])。
3)X 2 Is characterized by comprising the following steps: judging |X 2 -X 1 If I is less than standard1, if X 2 -X 1 |<Standard1, X 2 Normal value, otherwise X 2 Let X be an outlier 2 ＝mean([X 1 ,X 3 ])。
4)X 3 ,X 4 ,…,X N/I-1 Is characterized by comprising the following steps: judging |X i -X i-1 Whether or not is smaller than standard1, where i.e {3,4, …, N/I-1}, if |X i -X i-1 |<Standard1, X i Normal value, otherwise X i To be abnormal value, it is necessary to judge |X again i-1 -X i+1 Whether or not is less than standard2,if |X i-1 -X i+1 |<Standard2 is a discontinuous anomaly value, let X i ＝mean([X i-1 ,X i+1 ]) Otherwise X i ＝mean([X i-2 ,X i-1 ])。
5)X N/I Is characterized by comprising the following steps: judging |X N/I -X N/I-1 If I is less than standard1, if X N/I -X N/I-1 |<Standard1, X N/I Normal value, otherwise X N/I Let X be an outlier N/I ＝mean([X N/I-1 ,X N/I-2 ])。
Step 3, AR modeling and secondary feature extraction of centroid sequence
The outlier-removed centroid sequence of all targets was fitted using a centralized p-order Autoregressive (AR) model. Centroid sequence X after eliminating outliers 1 ,X 2 ,…,X N/I The result of the centering AR (p) model fitting is as follows:
X t ＝c 1 X t-1 +…+c p X t-p +ε t (9)
wherein the model order p is a known quantity, and the model parameter c= (c) 1 ,c 2 ,…,c p ) T For the parameters to be estimated, the moment estimates thereofCan be calculated by equation (10).
in the formula ,equation (10) is called a You Er-wok equation estimation method of parameters for the estimation value of the autocorrelation coefficient. White noise variance->Moment estimate +.>The following are provided:
in the formula ,variance of centroid feature sequence, ++>Is an autocovariance function of interval j.
The formula (9) is the AR model fitting result of a sea surface target centroid sequence, the AR model order p is generally set to be 3, and the estimated AR model coefficient isDenoted as AR (1), AR (2), AR (3), which are secondary features corresponding to the centroid sequence of a sea surface target.
Step 4, construction and identification of classifier
In the training stage of the classifier, extracting a large number of secondary features AR (1), AR (2) and AR (3) of Doppler spectrum centroids from the labeled ship target echo to form a required training set feature sample, and carrying out standard deviation normalization (Standard Deviation Normalization) on the training set feature sample; the error judgment probability of a ship target is preset, and training is carried out by using a rapid convex hull learning algorithm to obtain a judgment area, so that training of the classifier is completed. In the classifier identification stage, a target echo sample [ AR (1), AR (2), AR (3) ] acquired by a radar is input into a trained convex hull single classifier, and the target class is identified.
Compared with the prior art, the sea surface floating target identification method utilizing centroid time sequence information has the beneficial effects that:
(1) The method provided by the invention utilizes Doppler spectrum centroid time sequence evolution rule information of the floating target, improves the utilization rate of target echo information, and solves the problem that the target cannot be identified in a refined way only by relying on one-dimensional range profile characteristics of the target in the past.
(2) The invention provides a double-threshold method for eliminating abnormal values of a centroid sequence, which is suitable for the common abnormal value eliminating situation.
(3) On the second-level observation time, the identification accuracy of the method provided by the patent to the targets with large mass difference such as ship targets and common floating targets can reach more than 90%.
Drawings
FIG. 1 is a flow chart of an embodiment of the present invention;
FIG. 2 is a representative Doppler spectrum centroid sequence of a ship target and a floating target according to the present invention;
FIG. 3 is a typical Doppler spectrum centroid sequence of a ship target and a floating target before and after outlier rejection according to the present invention;
FIG. 4 shows the distribution of the ship targets and floating targets in the AR model coefficient domain of the centroid sequence;
fig. 5 is a schematic diagram of training and decision making of a convex hull classifier.
Detailed Description
For better understanding and implementation, a specific embodiment of the present invention is described in detail below with reference to the accompanying drawings.
The specific implementation flow of the target identification method is shown in fig. 1, and the processing flow of the invention is described in detail by referring to the attached drawings:
1) Extraction of Doppler spectrum centroid sequence
Assuming that sea surface targets are detected by the sea detection radar, a coherent pulse string x (N) with a length of N is received on a certain target distance unit, where n=1, 2, …, N. Cutting x (N), n=1, 2, …, N into non-overlapping short vectors u of length I according to formula (1) i 。
Each short vector u is calculated according to the formulas (2) to (8) i Extracting Doppler spectrum centroid characteristics, and marking the Doppler spectrum centroid sequence of the obtained sea surface target as X 1 ,X 2 ,…,X NI . Real worldThe Doppler spectrum centroid sequence of a typical ship target and a floating target (buoy) obtained by the measured data is shown in figure 2. Obviously, the Doppler spectrum centroid sequence of the ship target and the floating target has a change rule with larger difference, the centroid sequence of the ship target is generally concentrated at about 0Hz, and the ship target is stable and has smaller fluctuation; the centroid sequence of the floating target has obvious fluctuation and is approximate to a certain periodicity, so that the ship target and the floating target can be distinguished by utilizing the time sequence evolution rule information of the centroid sequence.
2) Outlier correction of centroid sequence
When the radar works, the Doppler spectrum centroid sequence extracted in the step 1) is often influenced by random interference such as co-frequency interference, noise, unknown source electromagnetic waves and the like, so that abnormal values often appear in the Doppler spectrum centroid sequence extracted in the step 1), and the double-threshold method is provided for correcting the Doppler spectrum centroid sequence extracted in the step 1), so that the Doppler spectrum centroid sequence is smoother, and the influence of the abnormal values is eliminated. Doppler spectrum centroid sequence X of sea surface target by adopting double-threshold method 1 ,X 2 ,…,X N/I The process flow is as follows:
1) Initializing: according to the approximate change trend of the centroid sequence, two decision threshold standards 1 and 2 are set. When the dimension of the centroid is the normalized frequency, the value of standard1 ranges from 0.03 to 0.15, and the value of standard2 ranges from 0.05 to 0.3, for example, standard 1=0.03 and standard 2=0.05.
2)X 1 Is characterized by comprising the following steps: judging |X 1 -mean([X 2 ,X 3 ]) Whether the absolute value is smaller than standard1, wherein mean (·) represents mean operation, and the absolute value is represented. If |X 1 -mean([X 2 ,X 3 ])|<Standard1, X 1 Normal value, otherwise X 1 Let X be an outlier 1 ＝mean([X 2 ,X 3 ])。
3)X 2 Is characterized by comprising the following steps: judging |X 2 -X 1 If I is less than standard1, if X 2 -X 1 |<Standard1, X 2 Normal value, otherwise X 2 Let X be an outlier 2 ＝mean([X 1 ,X 3 ])。
4)X 3 ,X 4 ,…,X N/I-1 Is characterized by comprising the following steps: judging |X i -X i-1 Whether or not is smaller than standard1, where i.e {3,4, …, N/I-1}, if |X i -X i-1 |<Standard1, X i Normal value, otherwise X i To be abnormal value, it is necessary to judge |X again i-1 -X i+1 If I is less than standard2, if X i-1 -X i+1 |<Standard2 is a discontinuous anomaly value, let X i ＝mean([X i-1 ,X i+1 ]) Otherwise X i ＝mean([X i-2 ,X i-1 ])。
5)X N/I Is characterized by comprising the following steps: judging |X N/I -X N/I-1 If I is less than standard1, if X N/I -X N/I-1 |<Standard1, X N/I Normal value, otherwise X N/I Let X be an outlier N/I ＝mean([X N/I-1 ,X N/I-2 ])。
The abnormal values of other common characteristic sequences can be corrected according to the flow, and Doppler spectrum centroid sequences of a typical ship target and a floating target (buoy) before and after the abnormal value correction obtained by actual measurement data are shown in the figure 3. Obviously, the double-threshold method provided by the patent can achieve a good outlier smoothing effect and can solve the problem of continuous outlier correction.
3) AR modeling and secondary feature extraction of centroid sequences
Centroid sequence X after outlier correction using a centralized AR (p) model 1 ,X 2 ,…,X N/I The fitting result is shown as a formula (9), and the model parameters are calculated by a You Er-Wook equation estimation method introduced by a formula (10).
The AR model order p is generally set to 3, and the estimated AR model coefficient isThe two-level characteristics are generally referred to as AR (1), AR (2) and AR (3), namely the secondary characteristics corresponding to the mass center sequence of a certain sea surface target.
The typical distribution of the ship target and floating target (buoy) obtained from the measured data in the AR model coefficient domain of the centroid sequence is shown in fig. 4. Obviously, the ship target and the floating target have good separability in the three-dimensional characteristic space formed by the AR (1), the AR (2) and the AR (3).
4) Construction and identification of classifier
After the secondary features AR (1), AR (2) and AR (3) corresponding to Doppler spectrum centroid sequences of different target echoes are extracted, the identification problem is the classification problem in the 3-dimensional feature space. The invention puts the problem under an abnormal detection frame, designs a single class classifier by utilizing a convex hull learning algorithm, and realizes the identification of ship targets and floating targets.
The single-class classifier can train by only depending on samples of one class of data, namely according to the labeled ship target echo feature vector set S, the single-class classifier can train in a 3-dimensional feature space according to the preset ship target error judgment probability P f Gradually training to obtain a ship target echo convex hull, thereby effectively solving the problem of unbalance of two types of samples. According to The minimum-Volume Criterion (The minimum-Volume Criterion), standard deviation normalization processing is needed to be carried out on The set S in advance, so that The influence of overlarge variation range of a certain dimension characteristic value on The recognition effect of other characteristics is avoided. Assume that the training set is expressed as:
S 0 ＝[AR(1) 0 ,AR(2) 0 ,AR(3) 0 ] (12)
AR (1) 0 ,AR(2) 0 ,AR(3) 0 The vector length is marked as Q, and the column vectors are respectively composed of values of secondary characteristics AR (1), AR (2) and AR (3) of mass center sequences of a large number of ship target units. The standard deviation of each feature component can be estimated as follows:
AR (1) q ,AR(2) q AR(3) q Respectively representing feature vectors AR (1) 0 ,AR(2) 0 ,AR(3) 0 And mean (·) represents the mean operator. The normalized sample set can be expressed as:
wherein AR (1), AR (2) and AR (3) respectively represent column vectors composed of corresponding normalized features.
Obviously, the ship target sample and the floating target sample should be gathered in different areas in the 3-dimensional feature space, so that in the training process, the closer the ship target sample forms a convex hull to the feature point of the floating target sample gathering area, the greater the possibility that the abnormal point is eliminated, and therefore, the classifier decision area is formed as follows:
1) Initializing: let the number of echo characteristic vectors of ship target be W, calculate the number of abnormal points F of ship target num ＝W·P f, wherein Pf The target error judgment probability of the ship is preset. Let l=0.
2) Finding the maximum value of AR (1) features in the set S, and forming a new space vertex v by the minimum values of AR (2) and AR (3) features 0 ＝[max(AR(1)),min(AR(2)),min(AR(3))]。
3) Generating a convex hull CH (S) of the current data point with a vertex { v } 1 ,v 2 ,…,v r }. Counting the number of characteristic points falling into the convex hull CH (S), and setting n as all 。
4) Calculating all feature points to the vertex v 0 Is found to be v 0 Furthest characteristic point v i And removing;
5) Generating a new convex hull CH (S- { v) i -n) then calculates the number of feature points in the new convex hull, set to n q 。
6) Let S- { v i }＝S,l+n all -n q ＝l。
7) If l<F num Returning to step 2) to continue with the removal of the next vertex. Otherwise, the removal process is terminated, and the final decision region Ω=ch (S) is output.
After the convex hull single classifier is trained, the collected target echo sample [ AR (1), AR (2), AR (3) ] is input, and the recognition can be completed, wherein the recognition rule is as follows:
the training and judging schematic diagram of the convex hull classifier obtained by actually measured data is shown in figure 5, wherein the centroid secondary characteristic of ship target echo is used as a sample, a convex hull single classifier is trained, and the preset ship target error judging probability P is obtained f 0% and a probability of correct identification of the floating target (buoy) of 98.09%.
Claims (2)
1. A floating target identification method using centroid time sequence information is characterized by comprising the following steps:
step 1, extraction of Doppler spectrum centroid sequence
Sea detection radar detects sea surface targets, and a coherent pulse string x (N) with the length of N is received on a certain target distance unit, wherein n=1, 2, … and N. Cut x (N), n=1, 2, …, N into non-overlapping short vectors u of length I i The following is shown:
[x(1),x(1),…,x(N)] T ＝[u 1 ,u 2 ,…,u N/I ] T (1)
respectively calculating each short vector u i The Doppler spectrum centroid characteristics are extracted, the power spectrum is used as the Doppler spectrum, the average Doppler spectrum and the short-time Doppler spectrum are estimated by adopting a non-parameter method in modern signal spectrum analysis, and windowing is adopted in the estimation to effectively reduce side lobes;
the Doppler spectrum estimation method comprises the following steps: assuming that the length of a single data segment is L, for the phi (phi=1, 2, …, phi) data segment, the short-time doppler spectrum estimate is expressed as:
wherein , and />Respectively representing a short-time magnitude spectrum and a short-time power spectrum, ω (n) representing a time-domain window function, T representing a pulse repetition period, c (φ) (n), n=1, 2, …, L represents the radar echo sequence of the phi-th data segment, delta is the window function power, expressed as:
and taking an average value of the short-time Doppler spectrum along the frequency dimension to obtain an average Doppler spectrum estimated value, wherein the average Doppler spectrum estimated value is expressed as:
doppler centroid f C Describing the Doppler shift degree of the target, the estimation method is as follows:
where Q (Φ) represents the power level of the short-term doppler spectrum for the Φ data segment, namely:
the Doppler spectrum centroid sequence of a certain sea surface target is finally obtainedDenoted as X 1 ,X 2 ,…,X N/I ；
Step 2, outlier correction of centroid sequence
Because of random interference such as co-frequency interference, noise, unknown source electromagnetic waves and the like, abnormal values often appear in the Doppler spectrum centroid sequence extracted in the step 1, a double-threshold method is provided, and the Doppler spectrum centroid sequence extracted in the step 1 is corrected to reduce the influence of the abnormal values;
step 3, AR modeling and secondary feature extraction of centroid sequence
The outlier-removed centroid sequence of all targets was fitted using a centralized p-order Autoregressive (AR) model. Centroid sequence X after eliminating outliers 1 ,X 2 ,…,X N/I The result of the centering AR (p) model fitting is as follows:
X t ＝c 1 X t-1 +…+c p X t-p +ε t (9)
wherein the model order p is a known quantity, and the model parameter c= (c) 1 ,c 2 ,…,c p ) T For the parameters to be estimated, the moment estimates thereofCan be calculated by formula (10);
in the formula ,equation (10) is called a You Er-wok equation estimation method of parameters for the estimation value of the autocorrelation coefficient. White noise varianceMoment estimate +.>The following are provided:
in the formula ,variance of centroid feature sequence, ++>An autocovariance function for interval j;
the formula (9) is the AR model fitting result of a sea surface target centroid sequence, the AR model order p is generally set to be 3, and the estimated AR model coefficient isThe two-level characteristics are marked as AR (1), AR (2) and AR (3), which are the secondary characteristics corresponding to the mass center sequence of a certain sea surface target;
step 4, construction and identification of classifier
In the training stage of the classifier, extracting a large number of secondary features AR (1), AR (2) and AR (3) of Doppler spectrum centroids from the labeled ship target echoes to form a required training set feature sample, and carrying out standard deviation normalization processing on the training set feature sample; the false decision probability of a ship target is preset, a quick convex hull learning algorithm is used for training, a decision area is obtained, training of a classifier is completed, a target echo sample [ AR (1), AR (2), AR (3) ] acquired by a radar is input into a convex hull single classifier which is already trained in a classifier identification stage, and the target class is identified.
2. A method for identifying a floating target using centroid timing information as defined in claim 1, wherein said double threshold method in step 2 is applied to a doppler spectrum centroid sequence X of a sea surface target 1 ,X 2 ,…,X N/I The process flow is as follows:
1) Initializing: setting two decision threshold standards 1 and 2 according to the approximate change trend of the centroid sequence, wherein when the dimension of the centroid is the normalized frequency, the value range of the standard1 is between 0.03 and 0.15, and the value range of the standard2 is between 0.05 and 0.3, for example, standard 1=0.03 and standard 2=0.05;
2)X 1 is characterized by comprising the following steps: judging |X 1 -mean([X 2 ,X 3 ]) Whether the absolute value is smaller than standard1, wherein mean (·) represents mean operation, and the absolute value is represented. If |X 1 -mean([X 2 ,X 3 ])|<Standard1, X 1 Normal value, otherwise X 1 Let X be an outlier 1 ＝mean([X 2 ,X 3 ])；
3)X 2 Is characterized by comprising the following steps: judging |X 2 -X 1 If I is less than standard1, if X 2 -X 1 |<Standard1, X 2 Normal value, otherwise X 2 Let X be an outlier 2 ＝mean([X 1 ,X 3 ])；
4)X 3 ,X 4 ,…,X N/I-1 Is characterized by comprising the following steps: judging |X i -X i-1 Whether or not is smaller than standard1, where i.e {3,4, …, N/I-1}, if |X i -X i-1 |<Standard1, X i Normal value, otherwise X i To be abnormal value, it is necessary to judge |X again i-1 -X i+1 If I is less than standard2, if X i-1 -X i+1 |<Standard2 is a discontinuous anomaly value, let X i ＝mean([X i-1 ,X i+1 ]) Otherwise X i ＝mean([X i-2 ,X i-1 ])；
5)X N/I Is characterized by comprising the following steps: judging |X N/I -X N/I-1 If I is less than standard1, if X N/I -X N/I-1 |<Standard1, X N/I Normal value, otherwise X N/I Let X be an outlier N/I ＝mean([X N/I-1 ,X N/I-2 ])。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310576670.7A CN116699617A (en) | 2023-05-19 | 2023-05-19 | Floating target identification method by utilizing centroid time sequence information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310576670.7A CN116699617A (en) | 2023-05-19 | 2023-05-19 | Floating target identification method by utilizing centroid time sequence information |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116699617A true CN116699617A (en) | 2023-09-05 |
Family
ID=87844293
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310576670.7A Pending CN116699617A (en) | 2023-05-19 | 2023-05-19 | Floating target identification method by utilizing centroid time sequence information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116699617A (en) |
-
2023
- 2023-05-19 CN CN202310576670.7A patent/CN116699617A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109031287B (en) | Human body respiration signal detection method based on fast-RCNN through-wall ultra-wideband radar | |
| CN104569948B (en) | Sub-band adaptive GLRT LTD detection methods under sea clutter background | |
| CN110852201A (en) | Pulse signal detection method based on multi-pulse envelope spectrum matching | |
| CN106500671B (en) | Method for determining sea water depth by decomposing laser radar waveform based on LM algorithm | |
| CN110501683B (en) | Sea-land clutter classification method based on four-dimensional data characteristics | |
| WO2023284698A1 (en) | Multi-target constant false alarm rate detection method based on deep neural network | |
| CN111830480B (en) | Radar sea clutter short-time spectrum characteristic parameter estimation method and system | |
| CN110749871B (en) | Parameter estimation method of dual-polarization weather radar | |
| CN111024207B (en) | Automatic detection and judgment method for vector hydrophone line spectrum | |
| CN109782251A (en) | A kind of slower-velocity target discrimination method after ocean clutter cancellation | |
| CN113534115B (en) | Sonar target high-precision line spectrum extraction method based on active and passive combined processing | |
| CN110070562A (en) | A kind of context-sensitive depth targets tracking | |
| CN112130142A (en) | Micro Doppler feature extraction method and system for complex moving target | |
| CN111707998B (en) | Sea surface floating small target detection method based on connected region characteristics | |
| CN110308442B (en) | GM-PHD target tracking method of phased array radar in strong clutter environment | |
| CN108318876A (en) | A method of estimating submarine target depth and distance using single hydrophone | |
| CN109782249B (en) | Two-target correlation time delay estimation algorithm | |
| CN105261028A (en) | Energy aggregation degree measure-based sea and land clutter scene segmentation method | |
| CN113640768B (en) | Low-resolution radar target identification method based on wavelet transformation | |
| CN114646935A (en) | Sea clutter distribution parameter estimation method | |
| CN108196238B (en) | Clutter map detection method based on adaptive matched filtering under Gaussian background | |
| CN110361723B (en) | Time-frequency feature extraction method for Doppler radar moving target | |
| CN116699617A (en) | Floating target identification method by utilizing centroid time sequence information | |
| CN109166132B (en) | Side-scan sonar image target identification method with variable initial distance symbolic function | |
| CN115508805B (en) | Classification method for ocean detection echo signals of airborne laser radar |
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 |