CN113434617A - Behavior automatic division method and system based on ship track and electronic equipment - Google Patents
Behavior automatic division method and system based on ship track and electronic equipment Download PDFInfo
- Publication number
- CN113434617A CN113434617A CN202110685011.8A CN202110685011A CN113434617A CN 113434617 A CN113434617 A CN 113434617A CN 202110685011 A CN202110685011 A CN 202110685011A CN 113434617 A CN113434617 A CN 113434617A
- Authority
- CN
- China
- Prior art keywords
- division
- behavior
- ship
- track
- cost
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/29—Geographical information databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Optimization (AREA)
- Computational Mathematics (AREA)
- Pure & Applied Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Mathematical Analysis (AREA)
- Artificial Intelligence (AREA)
- Remote Sensing (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Algebra (AREA)
- Software Systems (AREA)
- Navigation (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention discloses a behavior automatic partitioning method, a behavior automatic partitioning system and electronic equipment based on ship tracks. By designing the automatic ship behavior dividing method based on the multi-ship track, the invention can not only keep the complete motion information (including longitude and latitude, speed to ground, course to ground and the like) of the ship, but also expand along with the increase of the number of the ships and the air-space characteristics, simultaneously avoid manual parameter setting and adjustment, and realize the automatic behavior mode identification and division of the ship track.
Description
Technical Field
The invention belongs to the field of computer big data processing, and relates to a ship behavior automatic division method, a system and electronic equipment, in particular to a ship trajectory-based behavior automatic division method, a system and electronic equipment.
Background
During the navigation process of the ship, the sensors carried along with the ship can record the track data of the ship, and a great deal of valuable information can be mined from the data, so that the method is applied to the fields of shipping management, ship collision avoidance and the like. However, due to the influence of factors such as complex and variable navigation environment and equipment failure, the ship track often has the problems of low sampling rate, multiple noise points, uneven granularity, irregular change and the like, and the processing and analysis of the navigation track are difficult. Therefore, it is necessary to identify different behavior patterns existing in the ship sailing track, perform behavior division on the ship track, and reduce the complexity of the track while keeping the motion characteristics of the ship.
There are three general categories of approaches to the study of trajectory segmentation. One type divides the trajectory into segments with similar motion parameter values based on time interval, speed variation, heading bias, and other numerical information metrics. The second category considers the geometrical shape of the ship track, and divides the space network structure of the track by methods such as distance measurement and the like. And the third category is that the track semantic features of the ship stopping points are combined for division. However, these methods usually only focus on one of the motion parameters and the track space structure of the ship, and cannot keep complete ship motion process information, and the model parameters need to be manually adjusted. In addition, the traditional method focuses on behavior division of a single ship track, cannot distinguish common behavior patterns of multiple ships on the same custom route and specific personalized behaviors of part of the ships, and cannot adapt to the requirements of complex and variable navigation scenes.
Therefore, it is necessary to invent an automatic ship behavior division method, which can automatically identify and divide the ship behaviors and practically improve the intelligent level of ship trajectory analysis on the premise of fully mining the space geographic information of the ship trajectory on the same route, not limiting the motion characteristic number and not needing to manually adjust parameters.
Disclosure of Invention
In order to solve the problems, the invention provides a behavior automatic division method, a behavior automatic division system and an electronic device based on ship tracks on the basis of a hidden Markov model, wherein a behavior mode is identified from track data of a plurality of ships on the same route, and automatic behavior division is realized.
The method adopts the technical scheme that: a behavior automatic division method based on ship tracks comprises the following steps:
step 1: selecting trajectory data of n typical ships on the same customary route, preprocessing the trajectory data, and encoding space coordinates into low-dimensional representation; wherein n is a preset value;
the track data comprises a ship name, MMSI, longitude, latitude, speed to ground, heading of a bow, ship type, draught and a timestamp;
the preprocessing comprises the steps of unifying the recorded number of each ship track data into r, and regarding each ship track as track data on the same section of habitual navigation route from the time 0 to the time r; the characteristics of dimensionalities such as the ground speed, the ground heading, the bow heading and the like are screened, cleaned and integrated; finally obtaining the trajectory data T ═ T of n ships1,T2…TnWhere T isiFor the trajectory sequence of the ith ship, it is denoted as Ti＝{p1,p2…pr-each record p in the sequence is a d-dimensional vector;
step 2: simultaneously inputting trajectory data of n ships, identifying and dividing behavior patterns in a time dimension when the number of the behavior patterns is M, and iteratively updating model parameters through a division result until the division cost L is minimum;
and step 3: successively inputting trajectory data of n ships, identifying and dividing behavior patterns on individual dimensions when the number of the behavior patterns is M, and iteratively updating model parameters through division results until the division cost L is minimum;
and 4, step 4: comparing the division cost L in the step 2 with the division cost L in the step 3, selecting and storing smaller model parameters and division results;
and 5: updating the behavior mode number M, subdividing the existing behavior modes based on a greedy method, and repeating the steps 2-4 until the mode number M which enables the division cost L to be minimum is determined, and obtaining a final division result.
The technical scheme adopted by the system of the invention is as follows: a behavior automatic partitioning system based on ship tracks comprises the following modules:
the module 1 is used for selecting and preprocessing the trajectory data of n typical ships on the same customary route, and coding the space coordinates into low-dimensional representation; wherein n is a preset value;
the track data comprises a ship name, MMSI, longitude, latitude, speed to ground, heading of a bow, ship type, draught and a timestamp;
the preprocessing comprises the steps of unifying the recorded number of each ship track data into r, and regarding each ship track as track data on the same section of habitual navigation route from the time 0 to the time r; the characteristics of dimensionalities such as the ground speed, the ground heading, the bow heading and the like are screened, cleaned and integrated; finally obtaining the trajectory data T ═ T of n ships1,T2…TnWhere T isiFor the trajectory sequence of the ith ship, it is denoted as Ti＝{p1,p2…pr-each record p in the sequence is a d-dimensional vector;
the module 2 is used for inputting the track data of n ships simultaneously, identifying and dividing behavior patterns in a time dimension when the number of the behavior patterns is M, and iteratively updating model parameters through division results until the division cost L is minimum;
the module 3 is used for inputting trajectory data of n ships in sequence, identifying and dividing behavior patterns on individual dimensions when the number of the behavior patterns is M, and iteratively updating model parameters through division results until the division cost L is minimum;
the module 4 is used for comparing the division cost L of the module 2 with that of the module 3, selecting and storing smaller model parameters and division results;
and the module 5 is used for updating the behavior mode number M, subdividing the existing behavior modes based on a greedy method, and repeatedly executing the modules 2 to 4 until the mode number M which enables the division cost L to be minimum is determined, and obtaining a final division result.
The technical scheme adopted by the electronic equipment is as follows: an electronic device, comprising: one or more processors; a storage device to store one or more programs that, when executed by the one or more processors, cause the one or more processors to implement a method for automatic partitioning of behavior based on vessel trajectories.
The ship behavior automatic dividing method based on the multi-ship track has the advantages that through the design of the ship behavior automatic dividing method based on the multi-ship track, complete motion information (including longitude and latitude, speed to ground, course to ground and the like) of a ship is reserved, the ship behavior automatic dividing method can be expanded along with the increase of the number of the ships and the air-air characteristics, manual parameter setting and adjustment are avoided, and automatic behavior mode identification and division of the ship track are achieved.
Drawings
FIG. 1 is a flow chart of a method according to an embodiment of the present invention.
Detailed Description
In order to facilitate the understanding and implementation of the present invention for those of ordinary skill in the art, the present invention is further described in detail with reference to the accompanying drawings and examples, it is to be understood that the embodiments described herein are merely illustrative and explanatory of the present invention and are not restrictive thereof.
In the field of machine learning, Hidden Markov Models (HMMs) are widely used in pattern recognition as a classical probabilistic model describing statistical characteristics of random processes. The HMM can learn a state transition process by converting a problem into a series of hidden states into an observation sequence and finding a state sequence with the highest probability under the condition that the observation sequence is known. The model parameters can also be estimated from the observed variables and the hidden variables in turn. The ship track implies the behavior change of the ship, and is a typical Markov process, and hidden behavior patterns and conversion among the patterns can be found from the ship track sequence based on the HMM theory, so that the division is further realized.
Referring to fig. 1, the method for automatically dividing behaviors based on ship tracks provided by the invention comprises the following steps:
step 1: and selecting and preprocessing the trajectory data of n typical ships on the same customary route, and encoding the space coordinates into a low-dimensional representation by using an S2 algorithm.
The ship track recorded by the sensor comprises ship name, MMSI, longitude, latitude, speed to ground, heading of ship head, ship type, draught and time stamp. In order to realize the dimension reduction of the geographic spatial characteristics, a general spatial point indexing algorithm Google S2 is selected to encode the longitude and the latitude, the algorithm is based on a Hilbert curve, matrix projection is carried out on a spherical surface, secondary transformation correction is adopted, then coordinates are mapped to projection, and the spatial coordinates are encoded into low-dimensional representation. Meanwhile, the record number of each ship track data is unified into r, and each ship track can be regarded as track data on the same section of habitual navigation route from 0 moment to r moment. In addition, the pretreatment also screens, cleans and integrates features in other dimensions. Finally obtaining the trajectory data T ═ T of n ships1,T2…TnWhere T isiFor the trajectory sequence of the ith ship, it can be represented as Ti＝{p1,p2…prEach record p in the sequence is a d-dimensional vector.
Step 2: and simultaneously inputting the track data of n ships, identifying and dividing the behavior patterns in the time dimension when the number of the behavior patterns is M, and iteratively updating the model parameters through the division result until the division cost L is minimum.
The specific implementation comprises the following substeps:
step 2.1: assuming that there are M behavior patterns, the usage parameter is { lambda ] for the trajectory data of n ships at the same time1…λMDelta Hidden Markov Model (HMM) computes the Viterbi path to get the maximum probability of each hidden state in each behavior pattern at time t (t is for [0, r)])。
Step 2.2: candidate partition points are determined according to the probability of mode transition occurring at each time t. Assuming that there are only two behavior patterns of m1 and m2, the probability that the pattern m1 is converted to m2 or left in the pattern m1 at the time point t can be expressed as follows, and so on for a plurality of behavior patterns.
Wherein p ism1；k(t-1) represents the mode transition probability, π, of mode m1 in hidden state k at time point t-1m2；h、Bm2；hRespectively representing the initial probability distribution and the emission matrix of the mode m2 in the hidden state h, Am1；k、Bm1；kRespectively representing the state transition matrix and the emission matrix, Δ, of mode m1 in hidden state k12、Δ11Representing transition probability matrices for modes 1 through 2 and modes 1 through 1, respectively.
Step 2.3: selecting the division point with the highest probability from all the candidate division points as the best division point, and dividing the n tracks to obtain M sub-track segment sets { S }1,S2,…,SMIn which S isiAnd representing the sub track segment set corresponding to the ith behavior mode.
Step 2.4: based on the division result of the step 2.3, a Baum-Welch algorithm is used for estimating a parameter lambda of each behavior mode1…λMAnd a mode transition matrix delta between them.
Step 2.5: and repeating the steps from 2.1 to 2.4, carrying out trajectory division again by using the new model parameters, and updating the HMM parameters until the division cost L is minimum.
The partitioning cost L consists of two parts. L isDCalled the description cost, where q represents the total number of all track segments after partitioning. L iscCalled coding cost, where θ is a transition parameter of behavior patterns i to x, and P denotes a track segment sjProbability of belonging to behavioral pattern i.
L＝LD+LC
LD＝log*(M)+log*(n)+log*(d)+qlog(n)+qlog(M)
And 3, step 3: successively inputting the track data of n ships, identifying and dividing behavior patterns on individual dimensions when the number of the behavior patterns is M, and iteratively updating model parameters through division results until the division cost L is minimum.
The specific implementation comprises the following substeps:
step 3.1: assuming that M behavior patterns exist, inputting a d-dimension tensor p in track data T of a ship, and calculating a coding cost L after the d-dimension tensor p is added to each behavior patternC。
Step 3.2: selecting the behavior mode m with the minimum coding cost*Using its parameter { λ*Delta is taken as a segmentation basis to obtain a sub-track segment set S of the behavior pattern*。
Step 3.3: and successively executing 3.1-3.2 steps on each tensor in each ship track sequence, and updating the behavior mode and the corresponding division result set.
Step 3.4: based on the division result of the step 3.3, a Baum-Welch algorithm is used for estimating a parameter lambda of each behavior mode1…λMAnd a mode transition matrix delta between them.
Step 3.5: and repeating the steps 3.1 to 3.4, performing trajectory division again by using the new model parameters, and updating the HMM parameters until the division cost L is minimum, wherein the definition of the division cost is the same as that in the step 2.
And 4, step 4: and (4) comparing the division cost L of the step 2 with the division cost L of the step 3, selecting smaller model parameters and division results and storing the smaller model parameters and the division results.
And once selection is carried out during each iteration, and the division result of the step 2 or the step 3 is adopted, so that the model not only excavates the common behavior pattern of the ships on the same route, but also does not omit the specific behavior pattern of the individual ships. And storing the obtained behavior pattern and the divided sub-track segment set into a stack structure as candidates.
And 5, step 5: updating the behavior mode number M, re-dividing the existing behavior mode based on a greedy method, and repeating the steps 2-4 until the mode number M which enables the division cost L to be minimum is determined, and obtaining a final division result.
The specific implementation comprises the following substeps:
step 5.1: pop one behavior pattern m*And corresponding set S of sub-track segments*Dividing the obtained product, and repeating the steps 2 to 4.
Step 5.2: and calculating the division cost L after further division. If L is smaller, the new behavior pattern and the division result are pushed; otherwise, the current navigation mode and the segmentation result are kept and no further division is carried out.
Step 5.3: and repeating the steps from 5.1 to 5.2 until the stack is empty, and obtaining a final same-route multi-ship behavior mode and an automatic division result.
Based on the division result of the invention, the navigation behavior modes and the conversion of various ships in different typical complex water areas can be researched, and if the ship dodges from straight navigation to right rudder turning, the mode conversion point can be used as the reference of the rudder turning point, so that the ship can smoothly enter the next straight navigation route. Based on the invention, the researches such as ship behavior prediction, abnormal behavior mode monitoring and the like can be further developed, and collision avoidance routes under typical meeting scenes can be divided, so that a foundation is provided for practical applications such as shipping management, navigation risk perception and the like.
It should be understood that the above description of the preferred embodiments is given for clarity and not for any purpose of limitation, and that various changes, substitutions and alterations can be made herein without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (7)
1. A behavior automatic division method based on ship tracks is characterized by comprising the following steps:
step 1: selecting trajectory data of n typical ships on the same customary route, preprocessing the trajectory data, and encoding space coordinates into low-dimensional representation; wherein n is a preset value;
the track data comprises a ship name, MMSI, longitude, latitude, speed to ground, heading of a bow, ship type, draught and a timestamp;
the preprocessing comprises the steps of unifying the recorded number of each ship track data into r, and regarding each ship track as track data on the same section of habitual navigation route from the time 0 to the time r; the characteristics of the ground speed, the ground course and the bow course are screened, cleaned and integrated; finally obtaining the trajectory data T ═ T of n ships1，T2…TnWhere T isiFor the trajectory sequence of the ith ship, it is denoted as Ti＝{p1，p2…pr-each record p in the sequence is a d-dimensional vector;
step 2: simultaneously inputting trajectory data of n ships, identifying and dividing behavior patterns in a time dimension when the number of the behavior patterns is M, and iteratively updating model parameters through a division result until the division cost L is minimum;
and step 3: successively inputting trajectory data of n ships, identifying and dividing behavior patterns on individual dimensions when the number of the behavior patterns is M, and iteratively updating model parameters through division results until the division cost L is minimum;
and 4, step 4: comparing the division cost L in the step 2 with the division cost L in the step 3, selecting and storing smaller model parameters and division results;
and 5: updating the behavior mode number M, subdividing the existing behavior modes based on a greedy method, and repeating the steps 2-4 until the mode number M which enables the division cost L to be minimum is determined, and obtaining a final division result.
2. The automatic behavior partitioning method based on ship tracks according to claim 1, wherein: in the step 1, a space point index algorithm Google S2 is adopted to encode the longitude and the latitude, the algorithm is based on a Hilbert curve, matrix projection is carried out on a spherical surface, secondary transformation correction is adopted, then coordinates are mapped to projection, and space coordinates are encoded into low-dimensional representation.
3. The automatic behavior partitioning method based on ship tracks as claimed in claim 1, wherein the step 2 is implemented by the following sub-steps:
step 2.1: assuming that there are M behavior patterns, the usage parameter is { lambda ] for the trajectory data of n ships at the same time1…λMDelta, the hidden Markov model HMM calculates the Viterbi path to obtain the maximum probability of each hidden state in each behavior mode at time t, and t is in the range of 0, r]；
Step 2.2: determining candidate division points according to the probability of mode conversion occurring at each time t; assuming that there are only two behavior patterns of ml and m2, the probability of the pattern m1 switching to m2 or remaining in the pattern m1 at the time point t is expressed as follows:
wherein p ism1；k(t-1) represents the mode transition probability, π, of mode m1 in hidden state k at time point t-1m2；h、Bm2；hRespectively representing the initial probability distribution and the emission matrix of the mode m2 in the hidden state h, Am1；k、Bm1；kRespectively representing the state transition matrix and the emission matrix, Δ, of mode m1 in hidden state k12、Δ11Transition probability matrices representing modes 1 to 2 and modes 1 to 1, respectively;
the analogy of various behavior patterns;
step 2.3: selecting the division point with the highest probability from all the candidate division points as the best division point, and dividing the n tracks to obtain M sub-track segment sets { S }1，S2，...，SMIn which S isiRepresenting a sub-track segment set corresponding to the ith behavior mode;
step 2.4: based on the partitioning results of step 2.3, the parameter λ of each behavior pattern is estimated using the Baum-Welch algorithm1...λMAnd a mode transition matrix Δ therebetween;
step 2.5: repeating the step 2.1 to the step 2.4, carrying out trajectory division again by using the new model parameters, and updating the HMM parameters until the division cost L is minimum;
dividing cost L ═ LD+LC；
LD＝log*(M)+log*(n)+log*(d)+qlog(n)+qlog(M)
Wherein L isDCalled description cost, q represents the total number of all track segments after division; l iscCalled the coding cost, θi*Transition parameter, P(s), for behavior pattern i to xj| i) represents a track segment sjProbability of belonging to behavioral pattern i.
4. The automatic behavior partitioning method based on ship tracks as claimed in claim 3, wherein the specific implementation of step 3 comprises the following sub-steps:
step 3.1: assuming that M behavior patterns exist, inputting a d-dimension tensor p in track data T of a ship, and calculating a coding cost L after the d-dimension tensor p is added to each behavior patternC；
Step 3.2: selecting the behavior mode m with the minimum coding cost*Using its parameter { λ*Delta is taken as a segmentation basis to obtain a sub-track segment set S of the behavior pattern*；
Step 3.3: successively executing steps 3.1-3.2 for each tensor in each ship track sequence, and updating the behavior mode and the corresponding division result set;
step 3.4: based on the partitioning results of step 3.3, the parameter λ for each behavior pattern is estimated using the Baum-Welch algorithm1...λMAnd a mode transition matrix Δ therebetween;
step 3.5: and repeating the step 3.1 to the step 3.4, performing track division again by using the new model parameters, and updating the HMM parameters until the division cost L is minimum.
5. The automatic behavior partitioning method based on ship tracks according to any one of claims 1 to 4, wherein the specific implementation of step 5 comprises the following sub-steps:
step 5.1: pop one behavior pattern m*And corresponding set S of sub-track segments*Dividing the obtained product, and repeating the steps 2 to 4;
step 5.2: calculating the division cost L after further division; if L is smaller, the new behavior pattern and the division result are pushed; otherwise, the current navigation mode and the segmentation result are kept and no further division is carried out;
step 5.3: and (5.1) repeating the step 5.1 to the step 5.2 until the stack is empty, and obtaining a final same-route multi-ship behavior mode and an automatic division result.
6. A behavior automatic division system based on ship tracks is characterized by comprising the following modules:
the module 1 is used for selecting and preprocessing the trajectory data of n typical ships on the same customary route, and coding the space coordinates into low-dimensional representation; wherein n is a preset value;
the track data comprises a ship name, MMSI, longitude, latitude, speed to ground, heading of a bow, ship type, draught and a timestamp;
the preprocessing comprises the steps of unifying the recorded number of each ship track data into r, and regarding each ship track as track data on the same section of habitual navigation route from the time 0 to the time r; the characteristics of the ground speed, the ground course and the bow course are screened, cleaned and integrated; finally obtaining the trajectory data T ═ T of n ships1，T2…TnWhere T isiFor the trajectory sequence of the ith ship, it is denoted as Ti＝{p1，p2…pr-each record p in the sequence is a d-dimensional vector;
the module 2 is used for inputting the track data of n ships simultaneously, identifying and dividing behavior patterns in a time dimension when the number of the behavior patterns is M, and iteratively updating model parameters through division results until the division cost L is minimum;
the module 3 is used for inputting trajectory data of n ships in sequence, identifying and dividing behavior patterns on individual dimensions when the number of the behavior patterns is M, and iteratively updating model parameters through division results until the division cost L is minimum;
the module 4 is used for comparing the division cost L of the module 2 with that of the module 3, selecting and storing smaller model parameters and division results;
and the module 5 is used for updating the behavior mode number M, subdividing the existing behavior modes based on a greedy method, and repeatedly executing the modules 2 to 4 until the mode number M which enables the division cost L to be minimum is determined, and obtaining a final division result.
7. An electronic device, comprising:
one or more processors;
storage means for storing one or more programs which, when executed by the one or more processors, cause the one or more processors to implement the automatic division of behavior based on vessel trajectories method of any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110685011.8A CN113434617B (en) | 2021-06-21 | 2021-06-21 | Behavior automatic division method and system based on ship track and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110685011.8A CN113434617B (en) | 2021-06-21 | 2021-06-21 | Behavior automatic division method and system based on ship track and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113434617A true CN113434617A (en) | 2021-09-24 |
| CN113434617B CN113434617B (en) | 2022-05-13 |
Family
ID=77756782
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110685011.8A Active CN113434617B (en) | 2021-06-21 | 2021-06-21 | Behavior automatic division method and system based on ship track and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113434617B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108334905A (en) * | 2018-02-08 | 2018-07-27 | 中电科技（合肥）博微信息发展有限责任公司 | Ship action trail recognition methods |
| US20190370746A1 (en) * | 2018-06-05 | 2019-12-05 | Signal Ocean Ltd | Carrier path prediction based on dynamic input data |
| CN111694913A (en) * | 2020-06-05 | 2020-09-22 | 海南大学 | Ship AIS (automatic identification System) track clustering method and device based on convolution self-encoder |
| CN111783960A (en) * | 2020-07-09 | 2020-10-16 | 中国人民解放军国防科技大学 | Ship track prediction method and system based on automatic encoder and bidirectional LSTM |
| CN112132346A (en) * | 2020-09-24 | 2020-12-25 | 中国科学院空天信息创新研究院 | Ship navigation track prediction method based on ship type |
-
2021
- 2021-06-21 CN CN202110685011.8A patent/CN113434617B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108334905A (en) * | 2018-02-08 | 2018-07-27 | 中电科技（合肥）博微信息发展有限责任公司 | Ship action trail recognition methods |
| US20190370746A1 (en) * | 2018-06-05 | 2019-12-05 | Signal Ocean Ltd | Carrier path prediction based on dynamic input data |
| CN111694913A (en) * | 2020-06-05 | 2020-09-22 | 海南大学 | Ship AIS (automatic identification System) track clustering method and device based on convolution self-encoder |
| CN111783960A (en) * | 2020-07-09 | 2020-10-16 | 中国人民解放军国防科技大学 | Ship track prediction method and system based on automatic encoder and bidirectional LSTM |
| CN112132346A (en) * | 2020-09-24 | 2020-12-25 | 中国科学院空天信息创新研究院 | Ship navigation track prediction method based on ship type |
Non-Patent Citations (1)
| Title |
|---|
| 刘炯炯、刘克中、杨星、熊勇、袁志涛、吴晓烈: "《基于规则的多船协调避碰决策算法及仿真验证》", 《中国航海》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113434617B (en) | 2022-05-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113780395B (en) | Mass high-dimensional AIS trajectory data clustering method | |
| Gao et al. | Ship-handling behavior pattern recognition using AIS sub-trajectory clustering analysis based on the T-SNE and spectral clustering algorithms | |
| Zhao et al. | Where are you heading? dynamic trajectory prediction with expert goal examples | |
| CN109191922B (en) | Large-scale four-dimensional track dynamic prediction method and device | |
| CN102799900B (en) | Target tracking method based on supporting online clustering in detection | |
| CN110046378B (en) | Selective hierarchical integration Gaussian process regression soft measurement modeling method based on evolutionary multi-objective optimization | |
| CN112906858A (en) | Real-time prediction method for ship motion trail | |
| CN108985380B (en) | Point switch fault identification method based on cluster integration | |
| CN110909909A (en) | Short-term traffic flow prediction method based on deep learning and multi-layer spatiotemporal feature map | |
| CN112541944B (en) | Probability twin target tracking method and system based on conditional variational encoder | |
| CN115494865B (en) | Unmanned aerial vehicle cluster situation analysis method and medium based on space-time diagram convolutional network | |
| CN113326731A (en) | Cross-domain pedestrian re-identification algorithm based on momentum network guidance | |
| CN111694913B (en) | Ship AIS track clustering method and device based on convolution self-encoder | |
| CN110795522B (en) | Method and device for predicting track position of mobile user | |
| CN112258557B (en) | Visual tracking method based on space attention feature aggregation | |
| CN111858565B (en) | Multi-dimensional high-precision track intelligent prediction method based on line segment clustering | |
| CN114611663A (en) | Customized pedestrian trajectory prediction method based on online updating strategy | |
| CN114882293A (en) | Random forest and ship target classification method based on AIS data feature optimization | |
| CN117765258A (en) | Large-scale point cloud semantic segmentation method based on density self-adaption and attention mechanism | |
| CN113434617B (en) | Behavior automatic division method and system based on ship track and electronic equipment | |
| CN116541708A (en) | Ship track prediction method and system integrating data quality control and converter network | |
| Wang et al. | Dsmcl: Dual-level stochastic multiple choice learning for multi-modal trajectory prediction | |
| CN115578574A (en) | Three-dimensional point cloud completion method based on deep learning and topology perception | |
| Zhang et al. | SIMPL: A Simple and Efficient Multi-agent Motion Prediction Baseline for Autonomous Driving | |
| CN103646407B (en) | A kind of video target tracking method based on composition distance relation figure |
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 |