CN110458054A - A kind of polarimetric SAR image is berthed Ship Detection - Google Patents
A kind of polarimetric SAR image is berthed Ship Detection Download PDFInfo
- Publication number
- CN110458054A CN110458054A CN201910684484.9A CN201910684484A CN110458054A CN 110458054 A CN110458054 A CN 110458054A CN 201910684484 A CN201910684484 A CN 201910684484A CN 110458054 A CN110458054 A CN 110458054A
- Authority
- CN
- China
- Prior art keywords
- population
- sar image
- polarimetric sar
- liveness
- svm
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 27
- 238000000034 method Methods 0.000 claims abstract description 21
- 230000011218 segmentation Effects 0.000 claims abstract description 19
- 238000012360 testing method Methods 0.000 claims abstract description 13
- 230000010287 polarization Effects 0.000 claims abstract description 11
- 238000011156 evaluation Methods 0.000 claims abstract description 8
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 7
- 238000000605 extraction Methods 0.000 claims abstract description 6
- 238000007619 statistical method Methods 0.000 claims abstract description 4
- 239000002245 particle Substances 0.000 claims description 50
- 238000012706 support-vector machine Methods 0.000 claims description 24
- 238000012549 training Methods 0.000 claims description 8
- 238000005457 optimization Methods 0.000 claims description 7
- 238000007689 inspection Methods 0.000 claims description 2
- 230000006870 function Effects 0.000 description 6
- 230000008569 process Effects 0.000 description 6
- 238000003860 storage Methods 0.000 description 5
- 230000008859 change Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000002474 experimental method Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 230000002411 adverse Effects 0.000 description 1
- 230000033228 biological regulation Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000007123 defense Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
- G06F18/24133—Distances to prototypes
- G06F18/24137—Distances to cluster centroïds
- G06F18/2414—Smoothing the distance, e.g. radial basis function networks [RBFN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/13—Satellite images
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Astronomy & Astrophysics (AREA)
- Remote Sensing (AREA)
- Multimedia (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
Berth Ship Detection the invention discloses a kind of polarimetric SAR image, described method includes following steps: step 1: polarimetric SAR image is carried out Polarization target decomposition, and then carries out polarization characteristic extraction by input polarimetric SAR image；Step 2: extra large land segmentation is carried out to the image after feature extraction, extra large land segmentation template is obtained, land strong scattering part is removed, obtains result after extra large land segmentation；Step 3: according to the segmentation result of step 2, the polarimetric SAR image naval vessel that berths is detected using PAPSO-SVM；Step 4: and evaluation for statistical analysis to testing result, using quality factor as evaluation index.The method improve the high problems of vessel at anchor detection false alarm rate under condition of small sample, solve the problems, such as naval vessel detection difficult of berthing to a certain extent.
Description
Technical field
The invention belongs to technical field of remote sensing image processing, are related to a kind of based on population liveness population
(PopulationActivity Particle Swarm Optimization, PA-PSO) Support Vector Machines Optimized (SVM)
Polarimetric SAR image is berthed Ship Detection.
Background technique
Synthetic aperture radar (SAR) is a kind of advanced means of the acquisition remote sensing information of round-the-clock round-the-clock, and polarization SAR
Then more traditional SAR includes that more polarization informations have been obtained by way of different transmitting receptions containing the original of multichannel
Echo, the difference of reaction terrestrial object information that can be more careful, especially has very the differentiation of man-made target and natural target
Big advantage.Naval vessel detection is the pith of naval target detection, to national defense safety, marine transportation, marine economy development, port
Naval vessel search under mouth urban planning and adverse weather condition searches and rescues all tools and has very important significance.And for the naval vessel that berths
Detection is even more the heavy difficult point of research.Currently, it is more difficult to obtain the largely naval vessel sample that berths, be to go to sea or enter a port naval vessel into
The heavy difficult point of row detection and analysis.
Support vector machines (SVM) is a kind of more effective machine learning sample classification side under the conditions of this Limited Number
Method by will be mapped to higher dimensional space in the sample of lower dimensional space, and introducing penalty factor, so that sample is classified, for convenience
It calculates and is mapped using kernel function, the common preferable kernel function of effect is Radial basis kernel function (RBF), for RBF core
Function SVM, there are two parameter is critically important, i.e. penalty factor c and kernel functional parameter g, they determine SVM generalization ability and
The accuracy of classification.Therefore, the selection of penalty factor and kernel functional parameter is particularly significant.
Summary of the invention
In order to enable the selection result of penalty factor and kernel functional parameter be optimal SVM performance, the present invention provides
A kind of polarimetric SAR image based on PA-PSO optimization SVM is berthed Ship Detection.The method improve stop under condition of small sample
The high problem of ship detection false alarm rate is moored, solves the problems, such as naval vessel detection difficult of berthing to a certain extent.
The purpose of the present invention is what is be achieved through the following technical solutions:
A kind of polarimetric SAR image is berthed Ship Detection, is included the following steps:
Step 1: polarimetric SAR image is carried out Polarization target decomposition, and then carries out polarization characteristic by input polarimetric SAR image
It extracts；
Step 2: extra large land segmentation is carried out to the image after feature extraction, extra large land segmentation template is obtained, by land strong scattering portion
Point remove, it is after obtaining extra large land segmentation as a result, comprising training sample and test sample in result after the segmentation of the sea land；
Step 3: according to the segmentation result of step 2, PAPSO-SVM (the improved particle group optimizing of population liveness is utilized
Algorithm of support vector machine) the polarimetric SAR image naval vessel that berths is detected, the specific steps are as follows:
Step 3 A: according to the different rank population liveness (PA) of the different definition of population invariable number to be operated, if wait grasp
Making population invariable number is N, then population liveness rank is 2N, wherein the value of population invariable number N requires to be that basis wants innovatory algorithm
Parameter determine, improved parameter be it is several, then corresponding population just has several, for example, the present invention in PAPSO be applied to
Two parameters of SVM are adjusted, so population invariable number N=2；
Step 3 B: improving PSO (population) algorithm using population liveness, generates range with each population and exists
1-2NRandom number as population liveness, different degrees of exchange is carried out to the particle with different population liveness, is obtained
To PA-PSO algorithm, wherein exchange method is as follows:
Population invariable number is N, then particle active degree has 2NA rank generates random number and represents particle liveness, random number
It is 1-2N, in 1-2NIn a liveness, and it is divided into (N+1) layer, every layer represents the population number for participating in exchange；The liveness class of each layer
Other number isLabel representative in the upper right corner has several populations to be exchanged, and arrangement value represents
The particle combinations generated after exchange, then population liveness classificationIt represents particle in population to keep intact, without exchange；No
It is which population participates in exchange, such as the liveness classification of n-th layer should have that same liveness classification, which decides,Class, respectively
It isWherein 2n-1+ 1 equivalence, which respectively represents, selects n different populations, by
For mathematics permutation and combination principle it is found that different n populations of selecting is selected to swap from N number of population, one is sharedIn not
Same combination；
Step 3 C: parameter adjustment is carried out to SVM with PA-PSO algorithm, obtains PAPSO-SVM method, specific steps are such as
Under: two populations are first generated in range respectively respectively self solving for support vector machines parameter penalty factor c and RBF kernel function g, are made
For the initial value of solution；Each feature of training sample is input in SVM again, is fed back using error detection pixel number, no
Disconnected iteration obtains optimal parameter combination；
Step 3 D: in the SVM after test sample being input to parameter optimization, target is carried out using PAPSO-SVM method
Detection；
Step 4: and evaluation for statistical analysis to testing result, using quality factor as evaluation index.
Compared with the prior art, the present invention has the advantage that
1, the adjustment of vector machine parameter is supported using population, be capable of fast and stable detects naval vessel.
2, particle swarm algorithm is optimized using population liveness, avoids and is easy using particle swarm algorithm adjusting parameter
Obtain the defect of locally optimal solution.
Detailed description of the invention
Fig. 1 is algorithm overall flow figure；
Fig. 2 is a kind of volume scattering characteristic results of goal decomposition；
Fig. 3 is extra large land segmentation template；
Fig. 4 is the influence that different population liveness moves population；
Fig. 5 is a kind of form that particle is likely to occur during iv-th iteration；
The Pauli exploded view of Fig. 6 experimental data；
Fig. 7 is the binary map of PAPSO-SVM algorithm detection；
Fig. 8 is the result figure of PAPSO-SVM algorithm detection.
Specific embodiment
Further description of the technical solution of the present invention with reference to the accompanying drawing, and however, it is not limited to this, all to this
Inventive technique scheme is modified or replaced equivalently, and without departing from the spirit and scope of the technical solution of the present invention, should all be covered
Within the protection scope of the present invention.
The present invention provides a kind of polarimetric SAR images based on PA-PSO optimization SVM to berth Ship Detection, such as Fig. 1
Shown, the method specific implementation step is as follows:
Step 1: feature extraction is carried out to polarimetric SAR image.
Polarimetric SAR image is inputted, polarimetric SAR image is subjected to Polarization target decomposition, and then carry out polarization characteristic extraction.This
Invention is by taking the volume scattering feature that Freeman is decomposed as an example, and the result of goal decomposition is as shown in Fig. 2, but actually need not be confined to
This.
Step 2: extra large land segmentation is carried out to polarimetric SAR image.
Using the methods of morphology, extra large land segmentation is carried out to polarimetric SAR image, obtains extra large land segmentation template, land is strong
Scattered portion removes, and extra large land segmentation result is as shown in Figure 3.
Step 3: the polarimetric SAR image naval vessel that berths is detected using PAPSO-SVM.
Step 3 A: according to the different rank population liveness (PA) of the different definition of population invariable number to be operated, if wait grasp
Making population invariable number is N, then population liveness rank is 2N, the population liveness of different stage can make particle in population different
Exchange degree.It is illustrated in the present invention with population invariable number to be operated for 2.Population liveness has level Four at this time, including does not live
Jump rank 0 partly enlivens two kinds of rank (mark is respectively 1,2) and enlivens rank 3 entirely.
Step 3 B: liveness is carried out to population using population liveness and changes operation.Generate 2NA random number conduct
The liveness of population.It is as follows to set liveness rule change: when random number is 0, particle is presented as disabled state, and particle is protected
It is constant to hold original numerical value；When random number is 1 and 2, particle is presented as semiactive status, some particles and optimal solution or random production
Raw particle carries out exchange of values, and when random number is 3, particle is presented as that full active state, the particle in two populations carry out
Exchange.Specific exchange process is as follows:
First select first particle in two populations as first group of initial solution, two populations obtained with traversal are optimal
That group solution be used as solution to be exchanged.Determine whether particle swaps and exchange regulation according to population liveness numerical value.
Such as: assuming that c population is population 1, for corresponding solution in first position of particle, g population is population 2, corresponding
Solution is in second position of particle.Under original state, Q0In first particle value in first position storage c population, second position
Set first particle value in storage g population；Q1In the obtained c population optimal solution of first position storage traversal, second position
The g population optimal solution that storage traversal obtains.
If population liveness is 0, storage result is remained unchanged；If population liveness is 1, two particles, first position
The value set swaps；If population liveness is 2, the value of two particles, second position is swapped；If population liveness is
3, the value of two positions swaps simultaneously, is equivalent to Q0 and Q1It exchanges, at this time in general, the case where particle does not become
Change.
Again with first group of particle value in two populations, with the optimal solution that is generated in last iterative process as to be exchanged
Solution.Operating process is identical as previous step rule, and obtaining Q2 and Q3 is respectively to be changed according to what the population liveness being randomly generated obtained
The combination of optimal solution and first group of particle value during generation.For first group of particle, corresponding iteration optimal solution is set as grain
Q after son itself, i.e. first time iteration intersect2And Q3Middle result is identical.
Still with first particle in two populations, and a random number is generated in the range of respective solution as newly
Particle is ready for exchanging.Operating process is still identical as the first step, obtained Q4And Q5It is middle store respectively newly generated RANDOM SOLUTION and
The random combine of first particle in population.
Aforesaid operations process is as shown in figure 4, with Q0And Q1Generation for, according to different population liveness (PA), obtain
Different parameter combinations completes the improvement to PSO algorithm, obtains PA-PSO algorithm.
Step 3 C: parameter adjustment is carried out to SVM using PA-PAO algorithm, obtains PAPSO-SVM method.
Firstly, being trained the preparation of data, training process is special using the volume scattering component and Span of Polarization target decomposition
The mode combined is levied, is fed back with the classification results of training sample, the foundation as adjusting parameter.
Next, training data is input in improved SVM classifier.First population is initialized and obtains population
Optimal solution.Two populations are first generated in range respectively respectively self solving for penalty factor c and RBF kernel function g, as the initial of solution
Value.Using PA-PSO algorithm, Q is obtained0To Q5Six groups of particles set gradually into SVM parameter and detect to training sample, calculate
The fitness of testing result selects the smallest particle combinations of fitness as iteration optimal solution this time, will this time iteration
Optimal solution combination preserves the exchange for next-generation particle, and is combined with the optimal solution of this time iteration and update population
First group of particle value.By taking a kind of situation being wherein likely to occur as an example, it is assumed that the population liveness that every group of particle obtains all is 0,
That is for particle without exchange, obtained result is as shown in Figure 5.Above step is repeated, is iterated, is previously set until reaching
Stopping criterion for iteration obtains updated optimal population.
Updated optimal population is traversed, two parameters that every group of particle is SVM in population, using SVM to instruction
Practice sample to be detected, carry out the calculating of error detection number of picture elements, selects best one group of fitness as final supporting vector
The parameter of machine, is applied in the classification of test sample.
In the present invention, swap operation itself is unrelated with population invariable number to be operated, but need the population invariable number that exchanges be with
Population invariable number to be operated is related.As long as swap operation is i.e. as shown in figure 5, population needs to swap operation, Q0 and Q1 are handed over
It changes, Q2 is exchanged with Q3, and Q4 is exchanged with Q5, if the population does not need to exchange, this 6 particles of Q0-Q6 do not change.
In the present invention, defines population liveness and the combination variety of particle is enriched, continually introducing new explanation and reservation
While optimal solution, the combination between different parameters is intervened, and according to the population liveness being randomly generated, to greatest extent
Avoid the possibility for falling into local optimum.Detailed process is, generates random number and is used as population liveness, the population there are two only
In the case where, when definition random number is 0, particle keeps original numerical value constant, when random number is 1 or 2 in sequence, population 1 or kind
Particle in group 2 carries out exchange of values with optimal solution or the particle being randomly generated, and when random number is 3, population is entirely active, hands at this time
It changes and is equivalent to particle and exchanges, do not have an impact.Population liveness realize every group of particle respectively with population optimal solution, each iteration
Optimal solution and the optimal solution of particle is randomly generated into row stochastic swap operation, to increase the multiplicity of particle namely solved
Property, achieve the purpose that algorithm is avoided to fall into local optimum.
Step 3 D: the detection of target is carried out using the support vector machines after parameter optimization.
In SVM after test sample being input to parameter optimization, detected.
Step 4: and evaluation for statistical analysis to testing result, using quality factor as evaluation index.
Quality factor are defined as follows:
In formula: FOM indicates quality factor, NtrIndicate the target number correctly detected in testing result, NfaIndicate false-alarm
The number of target, NacIndicate the number of realistic objective.Quality factor are bigger, illustrate that detection effect is better.
Effect of the invention can be further described by following experiment:
1, experimental data
This experiment uses UAVSAR airborne system full polarimetric SAR data obtained, and using L-band, azimuth resolution is
0.6 meter, range resolution is 1.6 meters.PolSAR image size is 2700 × 2500, and corresponding Pauli figure is as shown in Figure 6.
2, experiment content and analysis
The result of the improved PSO-SVM of population liveness is more fine (population invariable number to be operated is 2), totally 36 ships, just
Really detect 34, missing inspection 2, false-alarm 3, final quality factor is 0.872.Due to some ship figures hull part greatly
It is low to scatter energy, easily occurs intermediate phenomenon of rupture when label, does not repeat statistics number here.The improved PSO- of population liveness
SVM experimental result is as shown in Figure 7 and Figure 8.
Claims (4)
- The Ship Detection 1. a kind of polarimetric SAR image is berthed, it is characterised in that described method includes following steps:Step 1: polarimetric SAR image is carried out Polarization target decomposition, and then carries out polarization characteristic and mention by input polarimetric SAR image It takes；Step 2: extra large land segmentation is carried out to the image after feature extraction, extra large land segmentation template is obtained, land strong scattering part is gone Fall, it is after obtaining the segmentation of extra large land as a result, comprising training sample and test sample in result after the segmentation of the sea land；Step 3: according to the segmentation result of step 2, the polarimetric SAR image naval vessel that berths is detected using PAPSO-SVM；Step 4: and evaluation for statistical analysis to testing result, using quality factor as evaluation index.
- The Ship Detection 2. polarimetric SAR image according to claim 1 is berthed, it is characterised in that the tool of the step 3 Steps are as follows for body:Step 3 A: according to the different rank population liveness PA of the different definition of population invariable number to be operated, if population to be operated Number is N, then population liveness rank is 2N；Step 3 B: improving PSO algorithm using population liveness, generates range in 1-2 to each populationNRandom number make For population liveness, the particle with different population liveness carries out different degrees of exchange, obtains PA-PSO algorithm；Step 3 C: parameter adjustment is carried out to SVM with PA-PSO algorithm, obtains PAPSO-SVM method；Step 3 D: in the SVM after test sample being input to parameter optimization, the inspection of target is carried out using PAPSO-SVM method It surveys.
- The Ship Detection 3. polarimetric SAR image according to claim 1 is berthed, it is characterised in that the step 3 C's Specific step is as follows:Two populations are first generated in range respectively respectively self solving for support vector machines parameter penalty factor c and RBF kernel function g, are made For the initial value of solution；Each feature of training sample is input in SVM again, is fed back using error detection pixel number, no Disconnected iteration obtains optimal parameter combination.
- The Ship Detection 4. polarimetric SAR image according to claim 1 is berthed, it is characterised in that the quality factor It is defined as follows:In formula: FOM indicates quality factor, NtrIndicate the target number correctly detected in testing result, NfaIndicate false-alarm targets Number, NacIndicate the number of realistic objective.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910684484.9A CN110458054B (en) | 2019-07-26 | 2019-07-26 | Detection method for ship berthing by polarized SAR image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910684484.9A CN110458054B (en) | 2019-07-26 | 2019-07-26 | Detection method for ship berthing by polarized SAR image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110458054A true CN110458054A (en) | 2019-11-15 |
| CN110458054B CN110458054B (en) | 2021-07-06 |
Family
ID=68483749
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910684484.9A Expired - Fee Related CN110458054B (en) | 2019-07-26 | 2019-07-26 | Detection method for ship berthing by polarized SAR image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110458054B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102005135A (en) * | 2010-12-09 | 2011-04-06 | 上海海事大学 | Genetic algorithm-based support vector regression shipping traffic flow prediction method |
| CN103971160A (en) * | 2014-05-05 | 2014-08-06 | 北京航空航天大学 | Particle swarm optimization method based on complex network |
| CN108052968A (en) * | 2017-12-08 | 2018-05-18 | 哈尔滨工程大学 | A kind of perception intrusion detection method of QSFLA-SVM |
| CN108737429A (en) * | 2018-05-24 | 2018-11-02 | 桂林电子科技大学 | A kind of network inbreak detection method |
-
2019
- 2019-07-26 CN CN201910684484.9A patent/CN110458054B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102005135A (en) * | 2010-12-09 | 2011-04-06 | 上海海事大学 | Genetic algorithm-based support vector regression shipping traffic flow prediction method |
| CN103971160A (en) * | 2014-05-05 | 2014-08-06 | 北京航空航天大学 | Particle swarm optimization method based on complex network |
| CN108052968A (en) * | 2017-12-08 | 2018-05-18 | 哈尔滨工程大学 | A kind of perception intrusion detection method of QSFLA-SVM |
| CN108737429A (en) * | 2018-05-24 | 2018-11-02 | 桂林电子科技大学 | A kind of network inbreak detection method |
Non-Patent Citations (5)
| Title |
|---|
| ABD. SAMAD HASANBASAR 等: "Opinion Mining of Movie Review using Hybrid Method of Support Vector Machine and Particle Swarm Optimization", 《PROCEDIA ENGINEERING》 * |
| FENG, LEIHUA 等: "Model Predictive Control of Duplex Inlet and Outlet Ball Mill System Based on Parameter Adaptive Particle Swarm Optimization", 《MATHEMATICAL PROBLEMS IN ENGINEERING》 * |
| JIE-SHENG WANG 等: "ANFIS Based Time Series Prediction Method of Bank Cash Flow Optimized by Adaptive Population Activity PSO Algorithm", 《INFORMATION》 * |
| 王晨逸: "PolSAR图像近海岸船舶检测方法研究", 《万方数据知识服务平台》 * |
| 王进 等: "ALPSO-SVM道路限速标志识别", 《计算机科学》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110458054B (en) | 2021-07-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhang et al. | ShipRSImageNet: A large-scale fine-grained dataset for ship detection in high-resolution optical remote sensing images | |
| CN107609601A (en) | A kind of ship seakeeping method based on multilayer convolutional neural networks | |
| CN112560671B (en) | Ship detection method based on rotary convolution neural network | |
| CN109213204B (en) | AUV (autonomous underwater vehicle) submarine target searching navigation system and method based on data driving | |
| Li et al. | Incorporation of AIS data-based machine learning into unsupervised route planning for maritime autonomous surface ships | |
| Zhao et al. | A novel similarity measure for clustering vessel trajectories based on dynamic time warping | |
| Zhou et al. | PVT-SAR: An arbitrarily oriented SAR ship detector with pyramid vision transformer | |
| CN113920443A (en) | Yoov 5-based remote sensing directed target detection method | |
| He et al. | Ship detection without sea-land segmentation for large-scale high-resolution optical satellite images | |
| CN116563726A (en) | Remote sensing image ship target detection method based on convolutional neural network | |
| Spiliopoulos et al. | Knowledge extraction from maritime spatiotemporal data: An evaluation of clustering algorithms on Big Data | |
| Hou et al. | A neural network based on consistency learning and adversarial learning for semisupervised synthetic aperture radar ship detection | |
| Charłampowicz | Maritime container terminal service quality in the face of COVID-19 outbreak | |
| Zhou et al. | Small vessel detection based on adaptive dual-polarimetric feature fusion and sea–land segmentation in SAR images | |
| Sun et al. | NSD‐SSD: a novel real‐time ship detector based on convolutional neural network in surveillance video | |
| CN110458054A (en) | A kind of polarimetric SAR image is berthed Ship Detection | |
| Zhang et al. | Ship detection and recognition in optical remote sensing images based on scale enhancement rotating cascade R-CNN networks | |
| Xu et al. | Exploring similarity in polarization: Contrastive learning with siamese networks for ship classification in sentinel-1 SAR images | |
| Liu et al. | OAB‐YOLOv5: One‐Anchor‐Based YOLOv5 for Rotated Object Detection in Remote Sensing Images | |
| CN117079120A (en) | Target recognition model optimization method based on improved GA algorithm | |
| CN113723182B (en) | SAR image ship detection method under training sample limited condition | |
| Hu et al. | Statistical analysis of massive AIS trajectories using Gaussian mixture models | |
| CN109031298A (en) | A kind of closed loop feedback determination method recognizing radar resource self adaptive imaging | |
| CN109871731A (en) | The method, apparatus and computer storage medium of ship detecting | |
| CN114663743A (en) | Ship target re-identification method, terminal equipment and storage medium |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20210706 |