CN114492571B - Ship track classification method based on similarity distance - Google Patents
Ship track classification method based on similarity distance Download PDFInfo
- Publication number
- CN114492571B CN114492571B CN202111572146.XA CN202111572146A CN114492571B CN 114492571 B CN114492571 B CN 114492571B CN 202111572146 A CN202111572146 A CN 202111572146A CN 114492571 B CN114492571 B CN 114492571B
- Authority
- CN
- China
- Prior art keywords
- ship
- distance
- similarity distance
- track
- track points
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 26
- 238000012549 training Methods 0.000 claims abstract description 8
- 238000012163 sequencing technique Methods 0.000 claims abstract description 5
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 230000002194 synthesizing effect Effects 0.000 abstract description 2
- 230000000694 effects Effects 0.000 description 4
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 239000003643 water by type Substances 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
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
- G06F18/24147—Distances to closest patterns, e.g. nearest neighbour classification
-
- 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
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a ship track classification method based on similarity distance, which comprises the steps of firstly recording track point data of a ship and selecting a reference ship; calculating the position similarity distance and the speed similarity distance between the reference ship and other ships, and then synthesizing the two to calculate the final similarity distance; replacing the similarity distance in the KNN with the final similarity distance for training to obtain the optimal super-parameters; sequencing all final similarity distances according to the numerical value, and counting the number of times of MMSI values of the ships corresponding to ship track points for generating the first K final similarity distance values; the vessel with the highest MMSI value is of the same class as the reference vessel. The method mainly aims at the problem that the traditional Euclidean distance is not suitable for classifying the marine ship track, and improves the accuracy and rationality of classifying the ship track.
Description
Technical Field
The invention belongs to the technical field of ships, and particularly relates to a ship track classification method.
Background
To assist maritime supervisory personnel in tracking ships and ensuring safe sailing, international Maritime Organization (IMO) requires vessels with a total tonnage of 300 or more, vessels carrying more than 500 total tons of cargo, and passenger vessels not in international waters to be equipped with Automatic Identification Systems (AIS). However, the real-time nature of the ship track received from the AIS is low and the time intervals are irregular. Meanwhile, AIS data may be occasionally lost due to the reliability of communication, which may cause the ship track update to be suspended. For the sear, the better the integrity of the ship data, the more sufficient the response space and time to avoid accidents. Therefore, it is necessary to classify the motion trajectories of the vessels to enhance the management of the vessels. In recent years, data analysis and predictive modeling have become an emerging topic of research.
However, the designed similarity distance in the KNN (k-nearest neighbors) algorithm is Euclidean distance, and the similarity distance calculation mode is not suitable for clustering among ship tracks. The reason is that the ship track point data contains a large number of influencing factors, and cannot be directly regarded as the distance between coordinate points. This results in poor performance of the KNN algorithm on classification of ship trajectories.
In summary, how to improve the practical effect of the ship track classification method has important research significance.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention provides a ship track classification method based on similarity distance, which comprises the steps of firstly recording track point data of a ship and selecting a reference ship; calculating the position similarity distance and the speed similarity distance between the reference ship and other ships, and then synthesizing the two to calculate the final similarity distance; replacing the similarity distance in the KNN with the final similarity distance for training to obtain the optimal super-parameters; sequencing all final similarity distances according to the numerical value, and counting the number of times of MMSI values of the ships corresponding to ship track points for generating the first K final similarity distance values; the vessel with the highest MMSI value is of the same class as the reference vessel. The method mainly aims at the problem that the traditional Euclidean distance is not suitable for classifying the marine ship track, and improves the accuracy and rationality of classifying the ship track.
The technical scheme adopted by the invention for solving the technical problems comprises the following steps:
step 1: recording the track point data of all ships in the delta t time period, preprocessing the track point data, and deleting outliers and null values; selecting a ship as a reference ship;
step 2: selecting a track point of a reference ship at a moment t in a delta t time period as a reference track point; calculating the position similarity distance between the reference track point and all other ship track points in the deltat time period by adopting the formula (1):
Distance 1 ＝R*arccos[sin(λ 1 )sin(λ 2 )+cos(λ 1 )cos(λ 2 )cos(l 1 -l 2 )] (1)
wherein R is the earth averageRadius, l 1 、l 2 Longitude, lambda of two track points respectively 1 、λ 2 Latitude of two track points respectively;
step 3: selecting a ship SOG value in a t period, and calculating the speed similarity distance between a reference track point and all other ship track points in a delta t period by adopting the method (2):
Distance 2 ＝v 1 -v 2 (2)
wherein v is 1 、v 2 The navigational speeds of the two track points are respectively;
step 4: step 2 and step 3 obtain the position similarity distance and speed similarity distance of all the ship track points to form a data set;
step 5: defining a final similarity distance:
D＝a*Distance 1 +(1-a)*Distance 2 (3)
wherein a is weight, a is [0,1];
step 6: determining an optimal super parameter by adopting a leave-one-out method;
optionally selecting 1 data set as test set and the rest data sets as training set; d is used for replacing the similarity distance in the KNN for training, so that optimal super parameters K and a are obtained;
step 7: bringing the value a into the formula (3), calculating the final similarity distances between the reference track point and all other ship track points in the delta t time period, and sequencing all the final similarity distances according to the numerical value;
step 8: selecting the first K values of the ordered final similarity distance list, and counting the number of times of MMSI values of the ship corresponding to ship track points for generating the first K final similarity distance values;
step 9: and returning the MMSI value of the ship with the largest occurrence number, and enabling the ship to be similar to the reference ship.
The beneficial effects of the invention are as follows:
the method mainly aims at the problem that the traditional Euclidean distance is not suitable for classifying the marine ship track, and improves the accuracy and rationality of classifying the ship track.
Drawings
FIG. 1 is a flow chart of the method of the present invention.
Fig. 2 is a diagram showing classification accuracy of the method under each parameter according to the embodiment of the present invention.
Fig. 3 is a diagram of classification accuracy versus effect provided by an embodiment of the present invention.
Detailed Description
The invention will be further described with reference to the drawings and examples.
According to actual environment and ship track data, the density of ships in the near-port sea area is high, and due to the fact that factors of mutual influence exist between ships, a similarity distance-based classification algorithm is suitable for the scene, but similarity distances such as Euclidean distance and Manhattan distance related to a traditional classification method are not suitable for classification of sea ship track points, and similarity between the track points cannot be measured well, so that the method is based on a KNN method, improves the similarity distance of the similarity distances, and classifies the ship track points.
As shown in fig. 1, a ship track classification method based on similarity distance includes the following steps:
step 1: recording the track point data of all ships in the delta t time period, preprocessing the track point data, and deleting outliers and null values; selecting a ship as a reference ship;
step 2: selecting a track point of a reference ship at a moment t in a delta t time period as a reference track point; calculating the position similarity distance between the reference track point and all other ship track points in the deltat time period by adopting the formula (1):
Distance 1 ＝R*arccos[sin(λ 1 )sin(λ 2 )+cos(λ 1 )cos(λ 2 )cos(l 1 -l 2 )] (1)
wherein R is the average radius of the earth, l 1 、l 2 Longitude, lambda of two track points respectively 1 、λ 2 Latitude of two track points respectively;
step 3: selecting a ship SOG value in a t period, and calculating the speed similarity distance between a reference track point and all other ship track points in a delta t period by adopting the method (2):
Distance 2 ＝v 1 -v 2 (2)
wherein v is 1 、v 2 The navigational speeds of the two track points are respectively;
step 4: step 2 and step 3 obtain the position similarity distance and speed similarity distance of all the ship track points to form a data set;
step 5: defining a final similarity distance:
D＝a*Distance 1 +(1-a)*Distance 2 (3)
wherein a is weight, a is [0,1];
step 6: determining an optimal super parameter by adopting a leave-one-out method;
optionally selecting 1 data set as test set and the rest data sets as training set; d is used for replacing the similarity distance in the KNN for training, so that optimal super parameters K and a are obtained;
step 7: bringing the value a into the formula (3), calculating the final similarity distances between the reference track point and all other ship track points in the delta t time period, and sequencing all the final similarity distances according to the numerical value;
step 8: selecting the first K values of the ordered final similarity distance list, and counting the number of times of MMSI values of the ship corresponding to ship track points for generating the first K final similarity distance values;
step 9: and returning the MMSI of the ship with the largest occurrence number, wherein the ship is similar to the reference ship.
FIG. 2 shows the classification accuracy of the method of the present invention when a takes different values.
FIG. 3 is a graph showing the classification accuracy versus effect of the method of the present invention and the conventional KNN method.
As can be seen from the figure, the method of the invention achieves better effects.
Claims (1)
1. The ship track classification method based on the similarity distance is characterized by comprising the following steps of:
step 1: recording the track point data of all ships in the delta t time period, preprocessing the track point data, and deleting outliers and null values; selecting a ship as a reference ship;
step 2: selecting a track point of a reference ship at a moment t in a delta t time period as a reference track point; calculating the position similarity distance between the reference track point and all other ship track points in the deltat time period by adopting the formula (1):
Distance 1 ＝R*arccos[sin(λ 1 )sin(λ 2 )+cos(λ 1 )cos(λ 2 )cos(l 1 -l 2 )] (1)
wherein R is the average radius of the earth, l 1 、l 2 Longitude, lambda of two track points respectively 1 、λ 2 Latitude of two track points respectively;
step 3: selecting a ship SOG value in a t period, and calculating the speed similarity distance between a reference track point and all other ship track points in a delta t period by adopting the method (2):
Distance 2 ＝v 1 -v 2 (2)
wherein v is 1 、v 2 The navigational speeds of the two track points are respectively;
step 4: step 2 and step 3 obtain the position similarity distance and speed similarity distance of all the ship track points to form a data set;
step 5: defining a final similarity distance:
D＝a*Distance 1 +(1-a)*Distance 2 (3)
wherein a is weight, a is [0,1];
step 6: determining an optimal super parameter by adopting a leave-one-out method;
optionally selecting 1 data set as test set and the rest data sets as training set; d is used for replacing the similarity distance in the KNN for training, so that optimal super parameters K and a are obtained;
step 7: bringing the value a into the formula (3), calculating the final similarity distances between the reference track point and all other ship track points in the delta t time period, and sequencing all the final similarity distances according to the numerical value;
step 8: selecting the first K values of the ordered final similarity distance list, and counting the number of times of MMSI values of the ship corresponding to ship track points for generating the first K final similarity distance values;
step 9: and returning the MMSI value of the ship with the largest occurrence number, and enabling the ship to be similar to the reference ship.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111572146.XA CN114492571B (en) | 2021-12-21 | 2021-12-21 | Ship track classification method based on similarity distance |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111572146.XA CN114492571B (en) | 2021-12-21 | 2021-12-21 | Ship track classification method based on similarity distance |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114492571A CN114492571A (en) | 2022-05-13 |
| CN114492571B true CN114492571B (en) | 2024-03-01 |
Family
ID=81493958
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111572146.XA Active CN114492571B (en) | 2021-12-21 | 2021-12-21 | Ship track classification method based on similarity distance |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114492571B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021017577A1 (en) * | 2019-07-29 | 2021-02-04 | 南京莱斯网信技术研究院有限公司 | Ship-type-spoofing detection method employing ensemble learning |
| CN113537386A (en) * | 2021-08-01 | 2021-10-22 | 大连海事大学 | Ship typical motion track self-adaptive mining method based on improved K-Medoids clustering |
-
2021
- 2021-12-21 CN CN202111572146.XA patent/CN114492571B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021017577A1 (en) * | 2019-07-29 | 2021-02-04 | 南京莱斯网信技术研究院有限公司 | Ship-type-spoofing detection method employing ensemble learning |
| CN113537386A (en) * | 2021-08-01 | 2021-10-22 | 大连海事大学 | Ship typical motion track self-adaptive mining method based on improved K-Medoids clustering |
Non-Patent Citations (2)
| Title |
|---|
| 基于AIS数据的船舶运动模式识别与应用;魏照坤;周康;魏明;史国友;;上海海事大学学报;20160630(02);全文 * |
| 基于KNN的船舶轨迹分类算法;刘磊;初秀民;蒋仲廉;钟诚;张代勇;;大连海事大学学报;20180815(03);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114492571A (en) | 2022-05-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109740494B (en) | Port identification method based on ship track big data | |
| CN113553682B (en) | Data-driven multi-level ship route network construction method | |
| CN109753068B (en) | Multi-USV group collaborative collision avoidance planning method considering communication situation | |
| CN103531043B (en) | Based on the boats and ships point-to-point hours underway evaluation method of course line coupling | |
| CN112906830B (en) | Automatic generation method of ship optimal route based on AIS big data | |
| CN113312438B (en) | Marine target position prediction method integrating route extraction and trend judgment | |
| CN110175633B (en) | Fishing boat track abnormity detection method based on Beidou satellite transmission data | |
| CN112699315B (en) | AIS data-based ship abnormal behavior detection method | |
| CN113537386B (en) | Ship typical motion trail self-adaptive mining method based on improved K-Medoids clustering | |
| Tang et al. | Detection of abnormal vessel behaviour based on probabilistic directed graph model | |
| CN115131393A (en) | Trajectory prediction method, collision detection method, apparatus, electronic device, and medium | |
| CN111240229B (en) | Fault information processing system based on intelligent ship berthing-off simulation test | |
| CN115050214B (en) | AIS data-based ship collision risk prediction method | |
| CN114492086B (en) | Intelligent detection method for single-mooring anchor position of ship | |
| CN117232520A (en) | Ship intelligent navigation system and navigation method suitable for offshore navigation | |
| Xu et al. | Trajectory clustering for SVR-based Time of Arrival estimation | |
| CN115457300A (en) | Ship abnormal behavior detection method based on distance measurement and isolation mechanism | |
| Xu et al. | Improved vessel trajectory prediction model based on stacked-bigrus | |
| CN114492571B (en) | Ship track classification method based on similarity distance | |
| Wang et al. | Classification of ship trajectories by using naive Bayesian algorithm | |
| WO2020161843A1 (en) | Ship behavior analyzing device, ship behavior analyzing method, and non-transitory computer-readable medium | |
| CN117312891A (en) | Track prediction method based on ship data | |
| CN110175406A (en) | Based on more ship following behavior analogy methods under the conditions of ship networking | |
| Zhou et al. | A real-time scene parsing network for autonomous maritime transportation | |
| CN114690772A (en) | Meeting time-space prediction method and system for sea area non-tracking navigation ship |
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 |