CN111694913A - Ship AIS (automatic identification System) track clustering method and device based on convolution self-encoder - Google Patents
Ship AIS (automatic identification System) track clustering method and device based on convolution self-encoder Download PDFInfo
- Publication number
- CN111694913A CN111694913A CN202010507856.3A CN202010507856A CN111694913A CN 111694913 A CN111694913 A CN 111694913A CN 202010507856 A CN202010507856 A CN 202010507856A CN 111694913 A CN111694913 A CN 111694913A
- Authority
- CN
- China
- Prior art keywords
- track
- ship
- sub
- characteristic
- encoder
- 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
- 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
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/284—Relational databases
- G06F16/285—Clustering or classification
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
Abstract
The invention relates to a ship AIS track clustering method and device based on a convolution self-encoder. The ship AIS track clustering method based on the convolution self-encoder comprises the following steps: acquiring a continuous track of a ship, and dividing the continuous track into a plurality of sub-tracks; extracting the characteristic engineering of the plurality of sub-tracks to obtain a sub-track characteristic matrix; inputting the sub-track feature matrix into a multi-feature fusion self-encoder to obtain a position feature vector, a speed feature vector and a course feature vector; splicing the position eigenvector, the speed eigenvector and the course eigenvector to obtain potential eigenvectors of the ship track; and carrying out track clustering operation on the extracted ship track characteristic vector to obtain a ship track clustering result. The method of the invention does not need to select the space-time trajectory measurement method according to the related data volume and the trajectory type, the calculation complexity, the noise and other influence factors, and does not need a similarity distance formula, thereby saving the calculation time and resources.
Description
Technical Field
The invention relates to the technical field of software engineering, in particular to a ship AIS track clustering method and device based on a convolution self-encoder.
Background
The satellite AIS (Automatic Identification System) is a ship positioning technology, AIS message information sent by a ship is received through a low-orbit satellite, and the satellite forwards the received and decoded AIS message information to a corresponding earth station, so that a land management organization can master related dynamic information of the ship, and the monitoring of the ship sailing in a high sea area is realized.
In order to improve the shipping capacity and efficiency of the ship, the prior art clusters the ship navigation track information data acquired from the AIS, so that a shipping scheme can be predicted according to a clustering result.
Track clustering algorithms are the key and basis for solving practical problems, and are widely used in many real-world applications, such as: anomaly detection, moving object behavior prediction, activity understanding, 3D structure reconstruction, traffic monitoring, and the like. Many researchers have proposed many trajectory clustering methods that use some trajectory similarity measurement methods to quantify trajectory similarity and then apply some classical clustering algorithms, such as K-means, gaussian mixture models, and density-based application space clustering, to perform a clustering analysis on the trajectories.
The traditional trajectory clustering method generally needs to select a space-time trajectory measurement method according to the related data volume and trajectory type, the calculation complexity, noise and other influence factors, and has the main problem that the selection of the optimal similarity measurement formula needs a large amount of prior knowledge and extensive experiments, so that a large amount of calculation resources and time are wasted.
Disclosure of Invention
Based on the above, the invention aims to provide a ship AIS (automatic identification system) track clustering method and device based on a convolution self-encoder, so that a space-time track similarity measurement formula does not need to be selected manually, and the similarity calculation deviation based on track distance caused by the traditional similarity measurement method can be avoided.
In a first aspect, an embodiment of the present application provides a ship AIS track clustering method based on a convolution self-encoder, including the following steps:
acquiring a continuous track of a ship, and dividing the continuous track into a plurality of sub-tracks;
extracting the characteristic engineering of the sub-tracks to obtain a sub-track characteristic matrix, wherein the ith row in the matrix represents the characteristic value of the ith sub-track;
inputting the sub-track feature matrix into a multi-feature fusion self-encoder, and performing feature extraction on the ship feature vector of the position, the speed and the course through the multi-feature fusion self-encoder to obtain a position feature vector, a speed feature vector and a course feature vector;
splicing the position characteristic vector, the speed characteristic vector and the course characteristic vector to obtain a potential characteristic vector of a ship track;
and carrying out track clustering operation on the extracted ship track characteristic vector to obtain a ship track clustering result.
Optionally, dividing the continuous track into a plurality of sub-tracks includes:
when the change of the longitude difference or the latitude difference of a certain track point relative to the previous track point in the continuous track is smaller than a set threshold value and the speed of the track point is close to 0, determining the track point as the starting point of a sub-track;
and determining the track point with the speed close to 0 as the terminal point of the sub-track, wherein the change of the longitude difference or the latitude difference of the next track relative to the previous track point in the continuous track is smaller than a set threshold value.
Optionally, extracting features of the plurality of sub-tracks to obtain a sub-track feature matrix, and further comprising:
normalizing the features in the sub-track feature matrix according to the following formula:
wherein, the feature to be normalized is shown, min and max respectively show the maximum value and the minimum value of the feature, and' show the feature obtained after normalization processing.
Optionally, after performing normalization operation on the features in the sub-trajectory feature matrix, the method further includes:
and (3) filling 0 in other track data by taking the number of the longest sub track points as a standard, so that the lengths of the ship track characteristic vectors are the same.
Optionally, the multi-feature fusion self-encoder includes:
the first convolution automatic encoder is used for learning position type characteristics of a ship track;
the second convolution automatic encoder is used for learning the speed type characteristics of the ship track;
and the third convolution automatic encoder is used for learning the course type characteristics of the ship track.
Optionally, the location type feature includes a longitude and a latitude;
the speed type features include speed and acceleration;
the heading type features include a heading and a turn rate.
Optionally, for each convolution automatic encoder, the structure includes:
the input layer is used for acquiring a characteristic matrix of characteristic data of each ship track, expanding the sub-track data into sub-track vectors and transmitting the sub-track vectors to the encoder part;
the encoder consists of a convolution layer and a pooling layer, and the activation function is set as a relu function and used for performing convolution operation on the sub-track vectors and outputting extracted ship track characteristics;
a decoder for implementing a deconvolution function using a combination of the convolutional layer and the upsampling layer, adjusting the low-dimensional feature vector to a high-dimensional feature vector, and transmitting the high-dimensional feature vector to the output layer;
and the output layer is used for converting the high-dimensional characteristic vector into the size of the original track characteristic vector so as to enable the latitude of the input data to be the same as that of the input data of the input layer.
Optionally, the similarity between the output layer data and the input layer data is measured by using the mean square error, and the formula is as follows:
whereinMSE represents the mean square errorMRepresenting real track data, predictedMRepresenting predicted trajectory data and N representing the total number of sub-trajectories for all vessels.
Optionally, the method further includes:
and (3) evaluating the clustering effect by adopting an f1 value method, wherein the formula is as follows:
where TP denotes a track predicted correctly as a positive example, FP denotes a track predicted incorrectly as a positive example, and FN denotes a track predicted incorrectly as a negative example.
In a second aspect, an embodiment of the present application provides a ship AIS track clustering device based on a convolution self-encoder, where the device includes:
the track acquisition module is used for acquiring a continuous track of a ship and dividing the continuous track into a plurality of sub-tracks;
the characteristic engineering module is used for extracting characteristic engineering of the sub-tracks to obtain a sub-track characteristic matrix, wherein the ith row in the matrix represents the characteristic value of the ith sub-track;
the characteristic extraction module is used for inputting the sub-track characteristic matrix into a multi-characteristic fusion self-encoder, and extracting the characteristics of the ship characteristic vector of the position, the speed and the course through the multi-characteristic fusion self-encoder to obtain a position characteristic vector, a speed characteristic vector and a course characteristic vector;
the splicing module is used for carrying out splicing operation on the position characteristic vector, the speed characteristic vector and the course characteristic vector to obtain a potential characteristic vector of a ship track;
and the clustering module is used for carrying out track clustering operation on the extracted ship track characteristic vector to obtain a ship track clustering result.
In the embodiment of the application, the path segmentation operation is carried out on the ship AIS original path to obtain the segmented ship AIS sub-paths, the characteristic engineering operation is carried out on each ship sub-path to obtain an AIS ship sub-path characteristic matrix, wherein the ith row in the matrix represents the characteristic value of the ith sub-path; the method comprises the steps of extracting the characteristics of ship characteristic vectors of position, speed and course by using a multi-characteristic fusion self-encoder to obtain three types of characteristic vectors, splicing the three types of characteristic vectors to obtain potential characteristic vectors of ship tracks, and carrying out track clustering operation on the extracted ship track characteristic vectors by using a mainstream clustering algorithm, so that a space-time track similarity measurement formula does not need to be selected manually, the similarity calculation deviation based on track distance caused by using a traditional similarity measurement method can be avoided, the distribution of data of different ship characteristic types can be reserved, and the calculation performance is improved.
For a better understanding and practice, the invention is described in detail below with reference to the accompanying drawings.
Drawings
FIG. 1 is a flowchart of a ship AIS trajectory clustering method based on a convolution self-encoder according to an embodiment of the present application;
FIG. 2 is a flowchart illustrating the steps of dividing a continuous track into sub-tracks according to an embodiment of the present application;
FIG. 3 is a schematic flow chart of a ship AIS trajectory clustering method based on a convolution self-encoder according to an embodiment of the present application;
FIG. 4 is a schematic structural diagram of a model applied to a ship AIS trajectory clustering method based on a convolution self-encoder according to an embodiment of the present application;
fig. 5 is a schematic structural diagram of a ship AIS track clustering device based on a convolution self-encoder in an embodiment of the present application.
Detailed Description
Reference will now be made in detail to the exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, like numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present specification. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the specification, as detailed in the appended claims.
The terminology used in the description herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the description. As used in this specification and the appended claims, the singular forms "a", "an", and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the term "and/or" as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items.
The following provides a detailed description of examples of the present specification.
Aiming at the technical problems that in the conventional technology, a spatiotemporal trajectory measurement method is generally selected according to related data volume and trajectory type, computational complexity, noise and other influence factors, and a large amount of computational resources and time are wasted, the embodiment of the application provides a ship AIS trajectory clustering method based on a convolution self-encoder, as shown in FIG. 1, and in one embodiment, the method comprises the following steps:
s101: acquiring a continuous track of a ship, and dividing the continuous track into a plurality of sub-tracks.
In the embodiment of the present application, first, a continuous track TR of a ship needs to be divided into several sub-tracks: TR { TR1, TR 2.., trj }, where trj is the ship sub-track and j represents the total number of sub-tracks for a ship.
The embodiment of the application divides the whole ship track TR into a series of ship sub-tracks, so that each ship sub-track can accurately represent a feasible and single sub-track from one port to another port, and longitude difference or latitude difference between two continuous points of the track needs to be calculated to identify the starting point and the ending point of the ship sub-track. Theoretically and empirically, the speed of a ship approaches 0 when it approaches a destination, and its latitude and longitude changes little.
In a specific example, as shown in fig. 2, dividing the continuous track into several sub-tracks includes the following steps:
s201: when the change of the longitude difference or the latitude difference of a certain track point relative to the previous track point in the continuous track is smaller than a set threshold value and the speed of the track point is close to 0, determining the track point as the starting point of a sub-track;
s202: and determining the track point with the speed close to 0 as the terminal point of the sub-track, wherein the change of the longitude difference or the latitude difference of the next track relative to the previous track point in the continuous track is smaller than a set threshold value.
S102: and extracting the characteristic engineering of the plurality of sub-tracks to obtain a sub-track characteristic matrix, wherein the ith row in the matrix represents the characteristic value of the ith sub-track.
In the embodiments of the present application, several concepts are first defined: given a data set a ═ { a 1, a 2, a 3, …, an }, where n is the number of vessels and ai is the time series data points of the sample points of a vessel trajectory. That is, ai is (p1, p 2, p 3.. pm), where m is the number of points in the sub-track, pm is defined as the point of the tuple (t, lat, lon, speed, cog, rot, ac), where t is the timestamp, lon is the longitude of the ship track point, lat is the latitude of the ship track point, cog is the heading of the ship track point, speed is the instantaneous speed of the ship track point, rot is the ship turning rate, and ac is the acceleration of the ship at time t.
Assuming that only the time, longitude and latitude of the track point can be obtained in some cases, but the feature extraction result can still be calculated from the raw data by using a ship track feature engineering formula as follows:
△latap＝latap-latap-1
△lonap＝lonap-lonap-1
△tap＝timeap-timeap-1
h＝hav(△latap)+cos(latap-1)*cos(latap)*hav(△lonap)
among them, △ latap，△lonap，△tapRespectively, a latitude difference, a longitude difference, and a time difference of the p-th point of the trajectory a. rot is the steering rate of the p point of the trajectory a, and Courseap is a piecewise function used for calculating the heading value of the p point of the trajectory a. h and haversin (theta) use the haversing formula to calculate the distance between two points between any two tracks on the earth, R is the earth radius, and the average is 6371 km. speed is the instantaneous speed of the track-off point calculated by dividing the distance between two points of the track by the time difference. acap represents the acceleration ac calculated from the speed difference and the time difference of two adjacent points of the trajectory.
In an embodiment, in order to scale different trajectory features to values between (0, 1) and enable the data extraction process to be faster and more accurate, the method further includes the following operations after performing feature engineering extraction on a plurality of sub-trajectories and obtaining a sub-trajectory feature matrix:
normalizing the features in the sub-track feature matrix according to the following formula:
wherein, the feature to be normalized is shown, min and max respectively show the maximum value and the minimum value of the feature, and' show the feature obtained after normalization processing.
In one embodiment, to simplify the training process, the method of converting the tracks of the ships with different lengths into fixed lengths, and after performing the normalization operation on the features in the sub-track feature matrix, further includes:
and (3) filling 0 in other track data by taking the number of the longest sub track points as a standard, so that the lengths of the ship track characteristic vectors are the same.
S103: and inputting the sub-track feature matrix into a multi-feature fusion self-encoder, and performing feature extraction on the ship feature vector of the position, the speed and the course through the multi-feature fusion self-encoder to obtain a position feature vector, a speed feature vector and a course feature vector.
The multi-feature fusion self-encoder is used for performing data dimension reduction on high-dimensional ship track feature data to obtain low-latitude ship track feature data.
Specifically, the multi-feature fusion self-encoder in the embodiment of the present application includes three convolution automatic encoders with the same structure, which are respectively:
the first convolution automatic encoder is used for learning position type characteristics of a ship track;
the second convolution automatic encoder is used for learning the speed type characteristics of the ship track;
and the third convolution automatic encoder is used for learning the course type characteristics of the ship track.
In one example, the location type features include longitude and latitude; the speed type features include speed and acceleration; the heading type features include a heading and a turn rate.
S104: and splicing the position characteristic vector, the speed characteristic vector and the course characteristic vector to obtain a potential characteristic vector of the ship track.
S105: and carrying out track clustering operation on the extracted ship track characteristic vector to obtain a ship track clustering result.
In the embodiment of the application, the three convolution self-encoders are used for feature extraction to obtain a ship track position class feature vector Z1, a ship track speed class feature vector Z2 and a ship track course class feature vector Z3, and finally the obtained three types of track feature vectors are spliced into a feature vector Z. And finally, carrying out track clustering operation on the extracted ship track characteristic vectors to obtain a ship track clustering result.
In one example, the AIS-based ship track low-dimensional feature vector Z can be used for clustering the track, and the clustering effect is evaluated by adopting an f1 value method, wherein the formula is as follows:
where TP denotes a track predicted correctly as a positive example, FP denotes a track predicted incorrectly as a positive example, and FN denotes a track predicted incorrectly as a negative example.
In the embodiment of the application, the path segmentation operation is carried out on the ship AIS original path to obtain the segmented ship AIS sub-paths, the characteristic engineering operation is carried out on each ship sub-path to obtain an AIS ship sub-path characteristic matrix, wherein the ith row in the matrix represents the characteristic value of the ith sub-path; the method comprises the steps of extracting the characteristics of ship characteristic vectors of position, speed and course by using a multi-characteristic fusion self-encoder to obtain three types of characteristic vectors, splicing the three types of characteristic vectors to obtain potential characteristic vectors of ship tracks, and carrying out track clustering operation on the extracted ship track characteristic vectors by using a mainstream clustering algorithm, so that a space-time track similarity measurement formula does not need to be selected manually, the similarity calculation deviation based on track distance caused by using a traditional similarity measurement method can be avoided, the distribution of data of different ship characteristic types can be reserved, and the calculation performance is improved.
In one embodiment, for each of the convolutional auto-encoders described above, the structure thereof comprises:
and the input layer is used for acquiring a characteristic matrix of characteristic data of each ship track, expanding the sub-track data into sub-track vectors and transmitting the sub-track vectors to the encoder part.
In one example, a feature matrix with dimensions (30, 2) of feature data of each ship track is obtained, where 30 represents that the number of ship sub-track points is also the length of the sub-track, and 2 represents each type of feature information included in a single track point, including: location, speed, or heading type characteristics. Meanwhile, in order to ensure the time sequence relation of the track sequence and make the convolution kernel filter completely cover all track points, the sub-track data is expanded into a sub-track vector with the length of 60 x 1 and is transmitted to the encoder part.
And the encoder consists of a convolution layer and a pooling layer, and the activation function is set as a relu function and is used for performing convolution operation on the sub-track vectors and outputting the extracted ship track characteristics.
In one example, the encoder is comprised of a convolutional layer and a pooling layer. Convolutional layer 1 consists of 4 (4, 1) convolutional kernels, followed by a pooling layer for downsampling, which uses maximal pooling for the pooling layer model. The activation function is set to the relu function and the step size is set to 2 because two trace points are convolved at a time so that the network structure of the convolutional layer arrangement can ensure that the time sequence relationship between the two trace points is kept. The ship track input of the model input is (60, 1), and the output dimensionality of the ship track feature extracted after convolution operation is (1, 4), wherein 1 represents 1 feature, and 4 represents 4 filters. The one-dimensional convolution formula is as follows:
where f and e represent the trajectory feature vector and the convolution kernel. α denotes the length of the feature vector and β denotes the convolution kernel size. The feature extraction of two track points can be carried out by the convolution kernel once convolution operation, the time sequence relation of the track points is guaranteed, local and context information about the ship track can be stored, and the analysis of the abnormal condition of the local ship track is facilitated.
And a decoder for implementing a deconvolution function using a combination of the convolutional layer and the upsampling layer, adjusting the low-dimensional feature vector into a high-dimensional feature vector, and transmitting the high-dimensional feature vector to the output layer.
In one example, in the decoder part, the deconvolution function is implemented by using a combination of convolutional layers and upsampling layers, which can ensure that the dimension of the output of the result after operation is not reduced, wherein convolutional layer 2 is composed of 4 (4, 1) convolutional kernel filters. The activation function is a relu function and the convolution step is set to 1. The effect of the final deconvolution is to adjust the low-dimensional feature vector z to a (60, 4) -dimensional feature vector and transmit it to the output layer.
And the output layer is used for converting the high-dimensional characteristic vector into the size of the original track characteristic vector so as to enable the latitude of the input data to be the same as that of the input data of the input layer.
In one example, the (60, 4) feature vector is converted into the original track feature vector size (60, 1), the output layer is composed of convolution kernels with the dimension of 1 (4, 1) and also a convolution layer, and the dimension adjustment of the output data is realized through convolution, so that the purpose of the same latitude as that of the input data of the input layer is finally realized.
In one example, to achieve quantization and evaluation of errors in the output-layer data and the input-layer data, the mean square error is used to measure the similarity between the output-layer data and the input-layer data, as follows:
where MSE represents the mean square error, observedMRepresenting real track data, predictedMRepresenting predicted trajectory data and N representing the total number of sub-trajectories for all vessels.
As shown in fig. 3 and 4, fig. 3 and 4 are a flowchart of a ship AIS track clustering method based on a convolutional auto-encoder and a model structure diagram applied in the embodiment of the present application.
In one example, the ship AIS track clustering method based on the convolutional auto-encoder of the present application first divides a ship continuous AIS track TR into several sub-tracks: TR { TR1, TR 2.., trj }, where trj is the ship sub-track, and j represents the total number of sub-tracks for a ship; performing characteristic engineering operation on each ship sub-track, normalizing the ship sub-tracks, and finally obtaining a normalized AIS ship sub-track characteristic matrix, wherein the ith row in the matrix represents the characteristic value of the ith sub-track; carrying out track filling on the matrix to ensure that the lengths of all sub-track vectors are equal, thereby facilitating the subsequent feature extraction operation; carrying out feature extraction on ship feature vectors of position, speed and course by using a multi-feature fusion self-encoder to obtain three types of feature vectors; and then splicing the three types of feature vectors to obtain the potential feature vectors of the ship track, and carrying out track clustering operation on the extracted ship track feature vectors by using a mainstream clustering algorithm.
Fig. 5 is a schematic structural diagram of a ship AIS track clustering device based on a convolution self-encoder according to an embodiment of the present application, and as shown in fig. 5, the ship AIS track clustering device 50 based on a convolution self-encoder includes:
a track obtaining module 51, configured to obtain a continuous track of a ship, and divide the continuous track into a plurality of sub-tracks;
the characteristic engineering module 52 is configured to perform characteristic engineering extraction on the plurality of sub-tracks to obtain a sub-track characteristic matrix, where an ith row in the matrix represents a characteristic value of an ith sub-track;
the characteristic extraction module 53 is configured to input the sub-trajectory characteristic matrix into a multi-characteristic fusion self-encoder, and perform characteristic extraction on the ship characteristic vector of the position, the speed and the course through the multi-characteristic fusion self-encoder to obtain a position characteristic vector, a speed characteristic vector and a course characteristic vector;
the splicing module 54 is configured to perform splicing operation on the position feature vector, the speed feature vector, and the heading feature vector to obtain a potential feature vector of a ship track;
and the clustering module 55 is used for performing track clustering operation on the extracted ship track characteristic vectors to obtain a ship track clustering result.
In an exemplary embodiment, the trajectory acquisition module 51 includes:
a starting point determining unit, configured to determine a track point as a starting point of a sub-track when a change of a longitude difference or a latitude difference of the track point relative to a previous track point in the continuous track is smaller than a set threshold and a speed of the track point is close to 0;
and the end point determining unit is used for determining a track point of which the change of the longitude difference or the latitude difference of the next track point relative to the previous track point in the continuous track is smaller than a set threshold and the speed is close to 0 as the end point of the sub-track.
In an exemplary embodiment, the vessel AIS track clustering apparatus 50 based on convolutional auto-encoder further includes:
the normalization module is used for performing feature engineering extraction on the sub-tracks to obtain a sub-track feature matrix, and then performing normalization operation on the features in the sub-track feature matrix according to the following formula:
wherein, the feature to be normalized is shown, min and max respectively show the maximum value and the minimum value of the feature, and' show the feature obtained after normalization processing.
In an exemplary embodiment, the vessel AIS track clustering apparatus 50 based on convolutional auto-encoder further includes:
and the track data filling unit is used for performing normalization operation on the features in the sub-track feature matrix, and then performing 0 filling on other track data by taking the number of the longest sub-track points as a standard, so that the lengths of the ship track feature vectors are the same.
In one exemplary embodiment, the multi-feature fusion self-encoder includes:
the first convolution automatic encoder is used for learning position type characteristics of a ship track;
the second convolution automatic encoder is used for learning the speed type characteristics of the ship track;
and the third convolution automatic encoder is used for learning the course type characteristics of the ship track.
In one exemplary embodiment, the location type characteristics include longitude and latitude;
the speed type features include speed and acceleration;
the heading type features include a heading and a turn rate.
In one exemplary embodiment, for each convolutional auto-encoder, the structure comprises:
the input layer is used for acquiring a characteristic matrix of characteristic data of each ship track, expanding the sub-track data into sub-track vectors and transmitting the sub-track vectors to the encoder part;
the encoder consists of a convolution layer and a pooling layer, and the activation function is set as a relu function and used for performing convolution operation on the sub-track vectors and outputting extracted ship track characteristics;
a decoder for implementing a deconvolution function using a combination of the convolutional layer and the upsampling layer, adjusting the low-dimensional feature vector to a high-dimensional feature vector, and transmitting the high-dimensional feature vector to the output layer;
and the output layer is used for converting the high-dimensional characteristic vector into the size of the original track characteristic vector so as to enable the latitude of the input data to be the same as that of the input data of the input layer.
In an exemplary embodiment, the vessel AIS track clustering apparatus 50 based on convolutional auto-encoder further includes:
and the similarity evaluation unit is used for measuring the similarity between the output layer data and the input layer data by using the mean square error, and the formula is as follows:
where MSE represents the mean square error, observedMRepresenting real track data, predictedMRepresenting predicted trajectory data and N representing the total number of sub-trajectories for all vessels.
In an exemplary embodiment, the vessel AIS track clustering apparatus 50 based on convolutional auto-encoder further includes:
the clustering effect evaluation unit is used for evaluating the clustering effect by adopting an f1 value method, and the formula is as follows:
where TP denotes a track predicted correctly as a positive example, FP denotes a track predicted incorrectly as a positive example, and FN denotes a track predicted incorrectly as a negative example.
It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described systems, apparatuses and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
The above-mentioned embodiments only express several embodiments of the present invention, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the inventive concept, which falls within the scope of the present invention.
Claims (10)
1. A ship AIS track clustering method based on a convolution self-encoder is characterized by comprising the following steps:
acquiring a continuous track of a ship, and dividing the continuous track into a plurality of sub-tracks;
extracting the characteristic engineering of the sub-tracks to obtain a sub-track characteristic matrix, wherein the ith row in the matrix represents the characteristic value of the ith sub-track;
inputting the sub-track feature matrix into a multi-feature fusion self-encoder, and performing feature extraction on the ship feature vector of the position, the speed and the course through the multi-feature fusion self-encoder to obtain a position feature vector, a speed feature vector and a course feature vector;
splicing the position characteristic vector, the speed characteristic vector and the course characteristic vector to obtain a potential characteristic vector of a ship track;
and carrying out track clustering operation on the extracted ship track characteristic vector to obtain a ship track clustering result.
2. The ship AIS track clustering method based on the convolutional auto-encoder according to claim 1, wherein dividing the continuous track into several sub-tracks comprises:
when the change of the longitude difference or the latitude difference of a certain track point relative to the previous track point in the continuous track is smaller than a set threshold value and the speed of the track point is close to 0, determining the track point as the starting point of a sub-track;
and determining the track point with the speed close to 0 as the terminal point of the sub-track, wherein the change of the longitude difference or the latitude difference of the next track relative to the previous track point in the continuous track is smaller than a set threshold value.
3. The ship AIS track clustering method based on the convolution self-encoder according to claim 1, wherein after feature engineering extraction is performed on a plurality of sub-tracks to obtain a sub-track feature matrix, the method further comprises:
normalizing the features in the sub-track feature matrix according to the following formula:
wherein, the feature to be normalized is shown, min and max respectively show the maximum value and the minimum value of the feature, and' show the feature obtained after normalization processing.
4. The ship AIS track clustering method based on the convolution self-encoder according to claim 3, wherein after the normalization operation is performed on the features in the sub-track feature matrix, the method further comprises:
and (3) filling 0 in other track data by taking the number of the longest sub track points as a standard, so that the lengths of the ship track characteristic vectors are the same.
5. The ship AIS trajectory clustering method based on the convolutional auto-encoder as claimed in claim 1, wherein the multi-feature fusion auto-encoder comprises:
the first convolution automatic encoder is used for learning position type characteristics of a ship track;
the second convolution automatic encoder is used for learning the speed type characteristics of the ship track;
and the third convolution automatic encoder is used for learning the course type characteristics of the ship track.
6. The ship AIS trajectory clustering method based on the convolutional auto-encoder as claimed in claim 5, wherein:
the location type characteristic comprises a longitude and a latitude;
the speed type features include speed and acceleration;
the heading type features include a heading and a turn rate.
7. The ship AIS trajectory clustering method based on the convolutional auto-encoder as claimed in claim 5, wherein:
for each convolutional auto-encoder, the structure includes:
the input layer is used for acquiring a characteristic matrix of characteristic data of each ship track, expanding the sub-track data into sub-track vectors and transmitting the sub-track vectors to the encoder part;
the encoder consists of a convolution layer and a pooling layer, and the activation function is set as a relu function and used for performing convolution operation on the sub-track vectors and outputting extracted ship track characteristics;
a decoder for implementing a deconvolution function using a combination of the convolutional layer and the upsampling layer, adjusting the low-dimensional feature vector to a high-dimensional feature vector, and transmitting the high-dimensional feature vector to the output layer;
and the output layer is used for converting the high-dimensional characteristic vector into the size of the original track characteristic vector so as to enable the latitude of the input data to be the same as that of the input data of the input layer.
8. The ship AIS trajectory clustering method based on the convolutional auto-encoder as claimed in claim 7, wherein:
the mean square error is used to measure the similarity between the output layer data and the input layer data, and the formula is as follows:
where MSE represents the mean square error, observedMRepresenting real track data, predictedMRepresenting predicted trajectory data and N representing the total number of sub-trajectories for all vessels.
9. The ship AIS trajectory clustering method based on the convolutional auto-encoder as claimed in claim 1, further comprising:
and (3) evaluating the clustering effect by adopting an f1 value method, wherein the formula is as follows:
where TP denotes a track predicted correctly as a positive example, FP denotes a track predicted incorrectly as a positive example, and FN denotes a track predicted incorrectly as a negative example.
10. A boats and ships AIS orbit clustering device based on convolution autoencoder, its characterized in that includes:
the track acquisition module is used for acquiring a continuous track of a ship and dividing the continuous track into a plurality of sub-tracks;
the characteristic engineering module is used for extracting characteristic engineering of the sub-tracks to obtain a sub-track characteristic matrix, wherein the ith row in the matrix represents the characteristic value of the ith sub-track;
the characteristic extraction module is used for inputting the sub-track characteristic matrix into a multi-characteristic fusion self-encoder, and extracting the characteristics of the ship characteristic vector of the position, the speed and the course through the multi-characteristic fusion self-encoder to obtain a position characteristic vector, a speed characteristic vector and a course characteristic vector;
the splicing module is used for carrying out splicing operation on the position characteristic vector, the speed characteristic vector and the course characteristic vector to obtain a potential characteristic vector of a ship track;
and the clustering module is used for carrying out track clustering operation on the extracted ship track characteristic vector to obtain a ship track clustering result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010507856.3A CN111694913B (en) | 2020-06-05 | 2020-06-05 | Ship AIS track clustering method and device based on convolution self-encoder |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010507856.3A CN111694913B (en) | 2020-06-05 | 2020-06-05 | Ship AIS track clustering method and device based on convolution self-encoder |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111694913A true CN111694913A (en) | 2020-09-22 |
| CN111694913B CN111694913B (en) | 2023-07-11 |
Family
ID=72479622
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010507856.3A Active CN111694913B (en) | 2020-06-05 | 2020-06-05 | Ship AIS track clustering method and device based on convolution self-encoder |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111694913B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113033443A (en) * | 2021-03-31 | 2021-06-25 | 同济大学 | Unmanned aerial vehicle-based automatic pedestrian crossing facility whole road network checking method |
| CN113221450A (en) * | 2021-04-27 | 2021-08-06 | 中国科学院国家空间科学中心 | Dead reckoning method and system for sparse and uneven time sequence data |
| CN113434617A (en) * | 2021-06-21 | 2021-09-24 | 武汉理工大学 | Behavior automatic division method and system based on ship track and electronic equipment |
| CN113535861A (en) * | 2021-07-16 | 2021-10-22 | 子亥科技(成都)有限公司 | Track prediction method for multi-scale feature fusion and adaptive clustering |
| CN113780395A (en) * | 2021-08-31 | 2021-12-10 | 西南电子技术研究所（中国电子科技集团公司第十研究所） | Mass high-dimensional AIS trajectory data clustering method |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070024494A1 (en) * | 2005-04-11 | 2007-02-01 | Dizaji Reza M | Classification system for radar and sonar applications |
| CN103456192A (en) * | 2013-09-01 | 2013-12-18 | 中国民航大学 | Terminal area prevailing traffic flow recognizing method based on track spectral clusters |
| CN105205145A (en) * | 2015-09-18 | 2015-12-30 | 中国科学院自动化研究所 | Track modeling and searching method |
| 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 |
| WO2019011298A1 (en) * | 2017-07-14 | 2019-01-17 | 腾讯科技（深圳）有限公司 | Target tracking method and apparatus, electronic device and storage medium |
| CN109447135A (en) * | 2018-10-12 | 2019-03-08 | 天津大学 | A kind of new ship method of trajectory clustering |
| CN109708638A (en) * | 2018-12-03 | 2019-05-03 | 江苏科技大学 | A kind of ship track point extracting method |
| US10345106B1 (en) * | 2015-10-29 | 2019-07-09 | National Technology & Engineering Solutions Of Sandia, Llc | Trajectory analysis with geometric features |
| CN110033051A (en) * | 2019-04-18 | 2019-07-19 | 杭州电子科技大学 | A kind of trawler behavior method of discrimination based on multistep cluster |
| CN110210352A (en) * | 2019-05-23 | 2019-09-06 | 中国人民解放军海军工程大学 | Ship track method for detecting abnormality based on navigation channel model |
| CN110232319A (en) * | 2019-05-07 | 2019-09-13 | 杭州电子科技大学 | A kind of ship Activity recognition method based on deep learning |
| CN110443287A (en) * | 2019-07-19 | 2019-11-12 | 北京航空航天大学 | A kind of mobile stream method for drafting of the crowd based on sparse track data |
-
2020
- 2020-06-05 CN CN202010507856.3A patent/CN111694913B/en active Active
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070024494A1 (en) * | 2005-04-11 | 2007-02-01 | Dizaji Reza M | Classification system for radar and sonar applications |
| CN103456192A (en) * | 2013-09-01 | 2013-12-18 | 中国民航大学 | Terminal area prevailing traffic flow recognizing method based on track spectral clusters |
| CN105205145A (en) * | 2015-09-18 | 2015-12-30 | 中国科学院自动化研究所 | Track modeling and searching method |
| US10345106B1 (en) * | 2015-10-29 | 2019-07-09 | National Technology & Engineering Solutions Of Sandia, Llc | Trajectory analysis with geometric features |
| WO2019011298A1 (en) * | 2017-07-14 | 2019-01-17 | 腾讯科技（深圳）有限公司 | Target tracking method and apparatus, electronic device and storage medium |
| 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 |
| CN109447135A (en) * | 2018-10-12 | 2019-03-08 | 天津大学 | A kind of new ship method of trajectory clustering |
| CN109708638A (en) * | 2018-12-03 | 2019-05-03 | 江苏科技大学 | A kind of ship track point extracting method |
| CN110033051A (en) * | 2019-04-18 | 2019-07-19 | 杭州电子科技大学 | A kind of trawler behavior method of discrimination based on multistep cluster |
| CN110232319A (en) * | 2019-05-07 | 2019-09-13 | 杭州电子科技大学 | A kind of ship Activity recognition method based on deep learning |
| CN110210352A (en) * | 2019-05-23 | 2019-09-06 | 中国人民解放军海军工程大学 | Ship track method for detecting abnormality based on navigation channel model |
| CN110443287A (en) * | 2019-07-19 | 2019-11-12 | 北京航空航天大学 | A kind of mobile stream method for drafting of the crowd based on sparse track data |
Non-Patent Citations (4)
| Title |
|---|
| JING CAO等: "PCA-based hierarchical clustering of AIS trajectories with automatic extraction of clusters" * |
| 彭祥文;高曙;初秀民;何阳;陆丛;: "基于Spark的船舶航行轨迹聚类方法", no. 03 * |
| 王江: "移动目标轨迹的时空模式挖掘方法研究" * |
| 袁冠等: "移动对象轨迹数据挖掘技术", vol. 1, 中国矿业大学出版社 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113033443A (en) * | 2021-03-31 | 2021-06-25 | 同济大学 | Unmanned aerial vehicle-based automatic pedestrian crossing facility whole road network checking method |
| CN113033443B (en) * | 2021-03-31 | 2022-10-14 | 同济大学 | Unmanned aerial vehicle-based automatic pedestrian crossing facility whole road network checking method |
| CN113221450A (en) * | 2021-04-27 | 2021-08-06 | 中国科学院国家空间科学中心 | Dead reckoning method and system for sparse and uneven time sequence data |
| CN113221450B (en) * | 2021-04-27 | 2024-03-12 | 中国科学院国家空间科学中心 | Space-time prediction method and system for sparse non-uniform time sequence data |
| CN113434617A (en) * | 2021-06-21 | 2021-09-24 | 武汉理工大学 | Behavior automatic division method and system based on ship track and electronic equipment |
| CN113535861A (en) * | 2021-07-16 | 2021-10-22 | 子亥科技(成都)有限公司 | Track prediction method for multi-scale feature fusion and adaptive clustering |
| CN113535861B (en) * | 2021-07-16 | 2023-08-11 | 子亥科技(成都)有限公司 | Track prediction method for multi-scale feature fusion and self-adaptive clustering |
| CN113780395A (en) * | 2021-08-31 | 2021-12-10 | 西南电子技术研究所（中国电子科技集团公司第十研究所） | Mass high-dimensional AIS trajectory data clustering method |
| CN113780395B (en) * | 2021-08-31 | 2023-02-03 | 西南电子技术研究所（中国电子科技集团公司第十研究所） | Mass high-dimensional AIS trajectory data clustering method |
| WO2023029461A1 (en) * | 2021-08-31 | 2023-03-09 | 西南电子技术研究所（中国电子科技集团公司第十研究所） | Massive high-dimensional ais trajectory data clustering method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111694913B (en) | 2023-07-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111694913B (en) | Ship AIS track clustering method and device based on convolution self-encoder | |
| CN113780395B (en) | Mass high-dimensional AIS trajectory data clustering method | |
| CN111160311A (en) | Yellow river ice semantic segmentation method based on multi-attention machine system double-flow fusion network | |
| Tu et al. | Modeling historical AIS data for vessel path prediction: A comprehensive treatment | |
| CN110018453B (en) | Intelligent model identification method based on aircraft track characteristics | |
| CN114022847A (en) | Intelligent agent trajectory prediction method, system, equipment and storage medium | |
| CN111142538B (en) | Unmanned ship operation strategy real-time decision-making method for water environment monitoring | |
| CN110933633B (en) | Onboard environment indoor positioning method based on CSI fingerprint feature migration | |
| CN112818969A (en) | Knowledge distillation-based face pose estimation method and system | |
| Perera et al. | Ship performance and navigation data compression and communication under autoencoder system architecture | |
| CN110673148A (en) | Active sonar target real-time track resolving method | |
| CN112465199A (en) | Airspace situation evaluation system | |
| CN114154619A (en) | Ship track prediction method based on CNN and BILSTM | |
| CN111367901B (en) | Ship data denoising method | |
| CN115512152A (en) | Ship track classification method and system combining CNN (CNN) neural network and LSTM neural network | |
| CN116608861A (en) | Ship track behavior abnormality detection method, system, device and storage medium | |
| CN110263836B (en) | Bad driving state identification method based on multi-feature convolutional neural network | |
| Xu et al. | Trajectory clustering for SVR-based Time of Arrival estimation | |
| Chen et al. | A TDV attention-based BiGRU network for AIS-based vessel trajectory prediction | |
| CN117372877A (en) | Star map identification method and device based on neural network and related medium | |
| CN114972918B (en) | Remote sensing image ship target identification method based on integrated learning and AIS data | |
| CN115755134A (en) | Vehicle positioning method and device based on Informer network and computer | |
| CN114706087A (en) | Underwater terrain matching and positioning method and system for three-dimensional imaging sonar point cloud | |
| CN112350733B (en) | Rapid track compression method based on effective distance | |
| CN114861759A (en) | Distributed training method of linear dynamic system model |
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 |