CN112015956A - Similarity determination method, device, equipment and storage medium for mobile object - Google Patents
Similarity determination method, device, equipment and storage medium for mobile object Download PDFInfo
- Publication number
- CN112015956A CN112015956A CN202010922759.0A CN202010922759A CN112015956A CN 112015956 A CN112015956 A CN 112015956A CN 202010922759 A CN202010922759 A CN 202010922759A CN 112015956 A CN112015956 A CN 112015956A
- Authority
- CN
- China
- Prior art keywords
- moving objects
- node
- relationship
- determining
- moving
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000012545 processing Methods 0.000 claims description 24
- 238000012549 training Methods 0.000 claims description 14
- 238000001914 filtration Methods 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 5
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000002203 pretreatment Methods 0.000 claims description 3
- 238000004422 calculation algorithm Methods 0.000 description 10
- 239000000523 sample Substances 0.000 description 8
- 238000010586 diagram Methods 0.000 description 7
- 238000005295 random walk Methods 0.000 description 5
- 238000010801 machine learning Methods 0.000 description 3
- 238000012806 monitoring device Methods 0.000 description 3
- 241001465754 Metazoa Species 0.000 description 2
- 230000006399 behavior Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000007726 management method Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 241000894007 species Species 0.000 description 1
- 239000000758 substrate Substances 0.000 description 1
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
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/901—Indexing; Data structures therefor; Storage structures
- G06F16/9024—Graphs; Linked lists
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/903—Querying
- G06F16/9035—Filtering based on additional data, e.g. user or group profiles
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/907—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/909—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using geographical or spatial information, e.g. location
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Library & Information Science (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
The application provides a similarity determination method, a similarity determination device, similarity determination equipment and a storage medium of a mobile object. The method comprises the following steps: determining a plurality of second moving objects with a same-row relationship in the plurality of first moving objects according to the track data of the plurality of first moving objects; establishing a relation graph according to a plurality of second moving objects with the same row relation; the relationship graph comprises nodes and edges; the node represents a second mobile object, and the edge represents that two connected nodes have a same-row relationship; determining a node sequence taking each second mobile object as an initial node according to the connection relation among the second mobile objects in the relation graph; determining the vector characteristics of each second moving object by using a preset model according to the node sequence; and determining the similarity of the second moving objects according to the vector characteristics of the second moving objects. According to the embodiment of the application, the relation between the mobile objects without the same line can be established based on the same line relation, so that the similarity between the mobile objects is determined, and the accuracy is high.
Description
Technical Field
The present application relates to the field of data processing technologies, and in particular, to a method, an apparatus, a device, and a storage medium for determining similarity of moving objects.
Background
With the advancement of mobile devices and identification technologies, a large amount of trajectory data is recorded. There are two main types of trace data sources, one from the external device: and moving object information data captured by the bayonet probe, which data records characteristics of the moving object. Another type of trajectory data is data generated by the moving object itself, such as positioning data generated by a moving device on a pedestrian, positioning data generated by a vehicle-mounted device, and position information including the moving object.
The similarity among a plurality of moving objects can be researched based on the track data, so that the behavior mode of the moving object hidden under the track data is analyzed, and the method can be applied to the fields of traffic scheduling, city planning, recommendation systems, privacy protection, public safety management and control and the like. For example, other mobile objects with similar behaviors to the target object, namely people with similar activity habits to the target object, are inquired through the track of the mobile object, so that a plurality of target objects are found; or, in a major public epidemic incident, other objects closely related to the mobile object are found, and important key clues are provided for tracing the spreading rule of the epidemic, searching the close contact of the patient with the personnel, managing and controlling the epidemic and the like. Therefore, it is a problem to be solved for those skilled in the art how to determine the similarity of moving objects more accurately.
Disclosure of Invention
The application provides a method, a device, equipment and a storage medium for determining similarity of moving objects.
In a first aspect, the present application provides a method for determining similarity of moving objects, including:
determining a plurality of second moving objects with a same-row relationship in the plurality of first moving objects according to the track data of the plurality of first moving objects;
establishing a relationship graph of the plurality of second moving objects according to the plurality of second moving objects with the same-row relationship; the relationship graph comprises a plurality of nodes and a plurality of edges; the node represents the second mobile object, and the edge represents that two connected nodes have a same-row relationship;
determining a node sequence taking each second mobile object as an initial node according to the connection relation among the second mobile objects in the relation graph;
determining the vector characteristics of each second moving object by using a first preset model according to a node sequence taking each second moving object as an initial node;
and determining the similarity of the second moving objects according to the vector characteristics of the second moving objects.
In a second aspect, the present application provides an apparatus for determining similarity of moving objects, comprising:
the determining module is used for determining a plurality of second moving objects with the same row relation in the plurality of first moving objects according to the track data of the plurality of first moving objects;
the processing module is used for establishing a relationship graph of the plurality of second moving objects according to the plurality of second moving objects with the same-row relationship; the relationship graph comprises a plurality of nodes and a plurality of edges; the node represents the second mobile object, and the edge represents that two connected nodes have a same-row relationship;
the processing module is further configured to determine a node sequence using each second mobile object as an initial node according to a connection relationship between each second mobile object in the relationship graph;
determining the vector characteristics of each second moving object by using a first preset model according to a node sequence taking each second moving object as an initial node;
the processing module is further configured to determine similarity of each second moving object according to the vector feature of each second moving object.
In a third aspect, the present application provides a computer-readable storage medium, on which a computer program is stored, and the computer program, when executed by a processor, implements the method of any one of the first aspect.
In a fourth aspect, an embodiment of the present application provides an electronic device, including:
a processor; and
a memory for storing executable instructions of the processor;
wherein the processor is configured to perform the method of any of the first aspects via execution of the executable instructions.
According to the method, the device, the equipment and the storage medium for determining the similarity of the mobile objects, a plurality of second mobile objects with the same row relation in a plurality of first mobile objects are determined according to the track data of the first mobile objects; establishing a relation graph according to a plurality of second moving objects with the same row relation; the relationship graph passes through nodes and edges; the node represents a second mobile object, and the edge represents that two connected nodes have a same-row relationship; according to the connection relation among the second moving objects in the relation graph, a node sequence with the second moving objects as initial nodes is determined, the node sequence with the second moving objects as the initial nodes represents the second moving objects, the connection among the moving objects which are not in the same row can be established, further, the vector characteristics of the second moving objects are determined by using a preset model according to the node sequence with the second moving objects as the initial nodes, the similarity of the second moving objects is further determined, and the accuracy is high.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and together with the description, serve to explain the principles of the disclosure.
Fig. 1 is a schematic view of an application scenario provided in an embodiment of the present application;
fig. 2 is a schematic flowchart of an embodiment of a method for determining similarity of moving objects provided in the present application;
FIG. 3 is a relational diagram of one embodiment of a method provided herein;
FIG. 4 is a schematic structural diagram of an embodiment of a mobile object similarity determination apparatus provided in the present application;
fig. 5 is a schematic structural diagram of an embodiment of an electronic device provided in the present application.
With the foregoing drawings in mind, certain embodiments of the disclosure have been shown and described in more detail below. These drawings and written description are not intended to limit the scope of the disclosed concepts in any way, but rather to illustrate the concepts of the disclosure to those skilled in the art by reference to specific embodiments.
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 implementations described in the exemplary embodiments below are not intended to represent all implementations consistent with the present disclosure. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present disclosure, as detailed in the appended claims.
The terms "comprising" and "having," and any variations thereof, in the description and claims of this application and the drawings described herein are intended to cover non-exclusive inclusions. For example, a process, method, system, article, or apparatus that comprises a list of steps or elements is not limited to only those steps or elements listed, but may alternatively include other steps or elements not listed, or inherent to such process, method, article, or apparatus.
First, a part of vocabulary and application scenarios related to the embodiments of the present application will be described.
The peer relationship refers to: and keeping a group of moving objects connected in density and moving together, and keeping a certain preset time length, so that the group of moving objects have the same row relation.
Density connection means that: for a group of moving objects O, if there is a sequence of objects O1,…onIn which oj,…oiIs a subset of the sequence, and two objects o before and after any of the sequenceiAnd oi+1Are directly density-connected, then called oiAnd ojAre connected with each other in density.
Direct density connection means: for theA group of moving objects O is given a distance threshold e and a density threshold mu, and N (O) is seti)＝{oj∈O|dist(oi,oj) E.ltoreq, if ojIn N (o)i) And N (o)i) Is greater than mu, then o is calledjAnd oiAre directly density connected.
Fig. 1 is a schematic view of an application scenario provided in an embodiment of the present application. As shown in fig. 1, the application scenarios of the embodiment of the present application may include, but are not limited to: electronic equipment 11, terminal equipment 12. The terminal device 12 includes, for example: user equipment such as cell-phones, panel computer, wearable equipment, vehicle-mounted terminal. The terminal device may collect trajectory data.
The electronic device 11 and the terminal device 12 may be connected to each other through a network.
Further, the method can also comprise the following steps: the monitoring device 13 may, for example, collect trajectory data via the monitoring device 13.
The method provided by the embodiment of the present application may be implemented by an electronic device, for example, a processor of the electronic device executing corresponding software codes, or may be implemented by the electronic device executing corresponding software codes and interacting data with the server 14.
In other scenarios, the method of the embodiment of the present application may also be executed by the server 14, which is not limited by the embodiment of the present application.
In the moving process, the mobile object may acquire trajectory data of the mobile object through a terminal device or a monitoring device, where the trajectory data includes a mobile object identifier ID, location information and time information, where the location information is, for example, location coordinates (such as Global Positioning System (GPS) coordinates or plane coordinates).
Wherein, the mobile object includes, for example: human, animal or vehicle, etc.
The trajectory data may find moving objects that are in the same row with a specific moving object, that is, moving objects having the same row relationship, for example, a plurality of moving objects have similar trajectories within a certain time period.
The similarity between the mobile objects can be determined based on the peer relationship between the mobile objects, for example, the rule of species migration is discovered through the research on the moving track of animals, the peer vehicle group is discovered through the research on the moving track of vehicles, and the method can be applied to the fields of traffic management, safety control and the like.
The technical idea of the method of the embodiment of the application is as follows:
according to the trajectory data, the same-row relation among the moving objects is established, based on the moving objects with the same-row relation, the relation among the moving objects without the same row can be established, the similarity among all the moving objects is further determined, and the accuracy is high.
The technical solution of the present application will be described in detail below with specific examples. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments.
Fig. 2 is a flowchart illustrating an embodiment of a method for determining similarity of moving objects according to the present application. As shown in fig. 2, the method provided by this embodiment includes:
Specifically, the trajectory data of the plurality of first moving objects is obtained, and the trajectory data includes a moving object identification ID, position information of the first moving object at a plurality of trajectory points, and time information.
The track data may be track data within a preset time period.
The plurality of second moving objects having the same row relationship may be that the second moving objects have the same or similar trajectories within a certain time period, for example, the second moving objects keep moving together with the density connection.
For example, for trajectory data of a plurality of first moving objects in a day, the plurality of first moving objects are divided into a plurality of parallel groups, {1,2,4}, {1,2, 3}, {5,6, 7}, {1, 6, 7}, and the like in a parallel relationship. The second moving objects in different same-row groups may be second moving objects having a same-row relationship in different time periods, for example, 9 am to 10 am, 1,2,4 have a same-row relationship, and 13 pm to 14 pm, 1, 6,7 have a same-row relationship.
Or, for the same time period, into different peer groups, {1,2,4}, {5,6, 7}, {9, 10, 11,12} etc., e.g., 9 am to 10 am, 1,2,4 have peer-to-peer relationships, 5,6,7 have peer-to-peer relationships, 9, 10, 11,12 have peer-to-peer relationships.
Based on the peer relationship, a relationship can be established between the second mobile objects without the peer relationship, for example, 1 and 5,6,7 may have a closer relationship.
Moving objects that are not in-line with other moving objects may subsequently be disregarded.
For example, 1000 moving objects, there are 800 moving objects divided into a plurality of parallel groups, and the remaining 200 moving objects are not considered in the subsequent consideration, i.e., the 200 moving objects are not considered for determining the similarity. The similarity is determined to be the similarity between the 800 moving objects.
102, establishing a relationship graph of a plurality of second moving objects according to the plurality of second moving objects with the same row relationship; the relationship graph comprises a plurality of nodes and a plurality of edges; the node represents a second mobile object, and the edge represents that two connected nodes have a same-row relationship;
specifically, the mobile objects are used as nodes, the same-row relationship among the mobile objects is used as an edge, a relationship graph is constructed, the relationship graph can represent the relationship among a plurality of second mobile objects, the relationship among the mobile objects without the same row can be established, the vector characteristics of each second mobile object are further obtained according to the relationship graph, and the accuracy of determining the similarity is high.
As shown in FIG. 3, node 25 has a peer relationship with node 32, node 12 has a peer relationship with node 1, and so on.
And 103, determining a node sequence taking each second mobile object as a starting node according to the connection relation among the second mobile objects in the relation graph.
Specifically, after the relationship graphs of the plurality of second moving objects are created, the node sequence using each second moving object as the start node is determined, for example, in fig. 3, the node sequence using the node 3 as the start node is {3, 14,8,2,20,31,34,15}, and the node sequence using the node 33 as the start node is {33, 24,28,9,34,21,30, 19}, etc.
In this embodiment, one or more node sequences may be selected from the node sequences using the same second mobile object as the start node.
The node sequence using each second moving object as an initial node can be obtained according to the edge of each node in the relational graph and according to a preset random walk strategy algorithm.
Wherein, the length of the node sequence may be a preset length.
In an embodiment, the length of the node sequence using each second mobile object as the starting node may be the same, which is not limited in this application.
By determining the node sequence using the second mobile object as the starting node, the second mobile object can be represented by the vector feature of the node sequence.
And step 104, determining the vector characteristics of each second moving object by using a first preset model according to the node sequence taking each second moving object as an initial node.
Specifically, the vector feature table of each second mobile object is, for example, a vector feature extracted from a semantic description of a node sequence in which the second mobile object is a starting node, and the semantic description of the second mobile object includes, for example, information of each node in the node sequence in which the second mobile object is the starting node.
For example, in fig. 3, the node sequence starting at the node 3 is {3, 14,8,2,20,31,34,15}, and the node 3 can be represented by a vector feature obtained by feature extraction of information of the node sequence.
The first preset model is, for example, a model established according to a machine learning algorithm, and is obtained by training parameters of the model through training data.
For example, a node sequence with each node as a starting node is taken as a text and used as a training corpus of a model established by a machine learning algorithm to train the model, and after the training is completed, a word vector feature of each node, that is, a vector feature corresponding to the node, can be obtained.
And 105, determining the similarity of the second moving objects according to the vector characteristics of the second moving objects.
The similarity can be characterized by similarity, the greater the similarity is, the more similar the moving objects are, and the smaller the similarity is, the less similar the moving objects are, and the smaller the relevance is.
Specifically, the similarity between the second moving objects is calculated through the vector characteristics of the second moving objects, and the similarity between the second moving objects is determined.
The similarity may be obtained by calculating cosine similarity or euclidean distance of the vector features of each second moving object, for example, calculating cosine of an included angle between the vector features or euclidean distance of the vector features.
In an embodiment, after determining the similarity of each second moving object, the following operations may be further performed:
and for any second moving object, determining a preset number of second moving objects similar to the second moving object.
Specifically, a preset number of second moving objects with a larger similarity may be selected according to the ranking result of the similarities between the second moving object and other second moving objects.
In the method of this embodiment, a plurality of second moving objects having a same-row relationship among a plurality of first moving objects are determined according to trajectory data of the plurality of first moving objects; establishing a relation graph according to a plurality of second moving objects with the same row relation; the relationship graph passes through nodes and edges; the node represents a second mobile object, and the edge represents that two connected nodes have a same-row relationship; according to the connection relation among the second moving objects in the relation graph, a node sequence with the second moving objects as initial nodes is determined, the node sequence with the second moving objects as the initial nodes represents the second moving objects, the connection among the moving objects which are not in the same row can be established, further, the vector characteristics of the second moving objects are determined by using a preset model according to the node sequence with the second moving objects as the initial nodes, the similarity of the second moving objects is further determined, and the accuracy is high.
The method of the embodiment of the application can be used for an online scene or an offline scene, and if the method is used for the online scene, the trajectory data is collected according to a certain time window; and if the scene is an off-line scene, slicing the acquired track data according to a certain time window size, wherein each time window contains the track data in the time period.
In an embodiment, step 101 may be preceded by the following operations:
preprocessing the track data to obtain preprocessed track data; the pre-treatment comprises at least one of: interpolation processing and filtering processing.
Specifically, when the trajectory data is acquired, the trajectory data of some trajectory points may be lost due to the device and the network signal, and therefore, the trajectory data may be preprocessed by interpolation or the like before the above scheme is executed.
According to the time interval between the track points in the track data, the track points are added and/or deleted, the time interval between two continuous track points in front and back is consistent, and no data or data sparseness in a certain time window is avoided.
Some track points in the track data may have a relatively long position deviation from other track points, that is, there is noise, which may be caused by an error occurring when the track data is acquired, and therefore, filtering processing needs to be performed on the track data, that is, smoothing processing is performed on the track points in the track data, for example, filtering processing is performed by methods such as median filtering, and the track points with relatively long deviation in the track data are removed.
On the basis of the above embodiment, step 101 may be implemented as follows:
clustering the track data according to a preset time window to obtain at least one cluster; the class cluster comprises a plurality of first moving objects; the trajectory data includes: the position information and the time information of the first moving object at a plurality of track points;
determining a plurality of second moving objects with the same row relation from a plurality of class clusters in a second preset time length; the second preset time length is greater than or equal to the time lengths of at least two preset time windows.
Specifically, for a group of moving objects moving together, the moving tracks of the moving objects are kept connected in a dense manner in space and are maintained for a period of time, so that the moving objects with the same row relationship can be found by a method of clustering and intersecting different time windows, and similar moving objects, namely a plurality of moving objects with greater similarity, can be further found based on the same row relationship, and the accuracy is higher.
And dividing the acquired track data according to preset time windows, wherein each preset time window corresponds to the track data of a plurality of track points in the time period. Clustering is carried out on the track data of a plurality of first moving objects in a time window to obtain one or more class clusters corresponding to the time window, wherein each class cluster can comprise a plurality of first moving objects.
The clustering can be performed according to the distance between the track points of different first moving objects, such as Euclidean distance and Manhattan distance.
For example, time window 1 corresponds to class cluster 1{1,2,4,8,14,20}, class cluster 2{5,6, 7, 11,17}, class cluster 3{19,21,27,30,33,34}, time window 2 corresponds to class cluster 4{1,2,4, 13, 18,22}, class cluster 5{1,5,6, 7, 11}, class cluster 6{16, 19,21,23,24, 27,28,29,30,33,34 }.
Further, the same first moving objects are searched from the cluster of the adjacent time windows, and the same first moving objects are marked as a plurality of second moving objects with the same row relation.
Specifically, the intersection of the clusters of the adjacent time windows may be taken to obtain the same first moving objects, and the same first moving objects form the candidate set in the same row.
For example, the intersection of the cluster 1 of the time window 1 and the cluster 4 of the time window 2 is used to obtain the candidate set of the same line as {1,2,4}, the intersection of the cluster 2 of the time window 1 and the cluster 5 of the time window 2 is used to obtain the candidate set of the same line as {5,6,7, 11}, and the intersection of the cluster 3 of the time window 1 and the cluster 7 of the time window 2 is used to obtain the candidate set of the same line as {19,21,27,30,33,34 }. {1,2,4} is a set of second mobile objects having a peer relationship, {5,6,7, 11} is a set of second mobile objects having a peer relationship, and {19,21,27,30,33,34} is a set of second mobile objects having a peer relationship.
The candidate set of the same row may continue to intersect with the cluster of the subsequent time window to obtain a final set or sets of second moving objects having a same row relationship, where each set includes a plurality of second moving objects.
In one embodiment, whether there is a peer relationship may be determined as follows.
If a plurality of second moving objects exist in the clusters of the first number of preset time windows within the second preset duration, and at least a second number of continuous preset time windows exist in the first number of preset time windows, determining that the plurality of second moving objects have a same-row relationship.
For example, the second preset duration includes 5 preset time windows, and if there are a plurality of second moving objects in each of the clusters of 3 preset time windows, and 2 of the 3 preset time windows are consecutive, that is, adjacent time windows.
For example, the cluster 1 of the time window 1, the cluster 3 of the time window 2, and the cluster 5 of the time window 4 all include {1,5,6,7,8,9, 15}, and the time window 1 and the time window 2 are adjacent time windows, so that the mobile objects {1,5,6,7,8,9, 15} are a group of second mobile objects having a same row relationship.
Further, in order to make the final calculation result more accurate, that is, a sufficient number of sample data is required, the number of the plurality of second moving objects having the collinear relationship is greater than or equal to the third number.
In the embodiment of the application, the peer-to-peer time length and the peer-to-peer times can be calculated while determining that the peer-to-peer relationship exists. The number of the same row refers to the number of the same row within a preset time period, for example, the number of the same row is 2 when the same row is in the morning and in the afternoon in one day.
For example, the peer-to-peer duration of {1,5,6,7,8,9, 15} is the duration of 4 time windows.
In the above embodiment, a plurality of second moving objects having a peer relationship are found by using a method of clustering and intersection taking with a locally continuous time constraint, so that the moving objects which are temporarily away are avoided from being missed to be judged as much as possible, the peer duration and the peer times between the moving objects are recorded, and finally, the trajectory similarity is determined according to the plurality of second moving objects having the peer relationship, so that the robustness is good.
In one embodiment of the present invention, the substrate is,
further, the relationship diagram further includes: the weight of the edge is determined according to the peer-to-peer duration and the peer-to-peer times of two nodes connected by the edge; the number of times of the same row is the number of times of the same row of the two nodes in a first preset time length.
For example, the first preset time period is one day, one week, one month, etc., and assuming that the first preset time period is one day, the two mobile objects are in the same row twice in one day, the one-time same row time period is 10 minutes, and the one-time same row time period is 20 minutes.
For example, if two moving objects exist in a certain cluster of two time windows, the in-line duration of the two moving objects may be the sum of the durations of the two time windows.
Specifically, the collinear durations and the collinear times of two moving objects having a collinear relationship may be weighted to obtain the weights of the edges corresponding to the two moving objects.
In an embodiment, step 103 may be specifically implemented as follows:
determining a node sequence taking each second moving object as an initial node by using a second preset model according to the connection relation among the second moving objects in the relation graph and the weight corresponding to the connection relation; the sequence of nodes includes at least one node;
the second preset model is obtained by training according to a plurality of sample data, and each sample data comprises: the relationship graph and the node sequence taking each node as a starting node in the relationship graph.
Specifically, after the relationship graphs of the plurality of second moving objects are established, a node sequence using each second moving object as a start node may be determined by using a second preset model obtained through pre-training according to the connection relationship between each node in the relationship graphs and the weight of each edge, for example, in fig. 3, the node sequence using the node 3 as the start node is {3, 14,8,2,20,31,34,15}, and the node sequence using the node 33 as the start node is {33, 24,28,9,34,21,30, 19}, etc.
In an embodiment, the second preset model is a model that is established based on a random walk strategy algorithm and obtained through training.
For example, the preset model may be a model established according to the node2vec algorithm, and the parameters of the model are trained through sample data.
And generating a node sequence taking each node as a starting node by utilizing a random walk strategy according to the edges of all nodes in the relational graph and the weights of the edges on the basis of a preset model established by a node2vec algorithm. And training the parameters of the model through a relation graph included by the sample data and a node sequence taking each node as an initial node in the relation graph to obtain the optimized model. The obtained node sequence taking each node as the starting node can accurately represent the characteristics of each node.
In an embodiment, further, according to a node sequence with each second moving object as a starting node, a vector feature of each second moving object is determined by using a first preset model. The first preset model can be established by adopting a word2vec algorithm and obtained by training. Taking the node sequence of each second moving object as an initial node as a text corpus corresponding to each second moving object, and obtaining the vector characteristics of each second moving object by using a first preset model.
And constructing a corpus in a word vector model through the node sequence, namely regarding the node sequence as a text as a training corpus of a word2vec algorithm model, training the model, and obtaining word vector representation of each node after training, namely vector characteristics of the node sequence taking the node as an initial node, namely vector characteristics of a second moving object corresponding to the node.
The word2vec algorithm is used to generate a correlation model of the word vector, and the model may be a neural network model. The word2vec model may map each word to a vector space of the same dimension, and the distances of similar words in the vector space are also similar.
It should be noted that, in other embodiments, the preset model may also be established by using other machine learning algorithms, which is not limited in the embodiment of the present application.
It should be noted that the first preset model and the second preset model may be used as one model or may be independent models, and the embodiment of the present application is not limited thereto.
In the above embodiment, based on the idea of node2vec, the weight of the edge is considered, a node sequence using each node as a starting node is generated by using a random walk strategy, so as to construct a corpus in a word vector model, and further obtain a vector representation of each node, which is used for finding similar moving objects, and allows no co-occurrence between the moving objects, so that the accuracy is high.
In the above embodiment, the relationship graph is constructed based on the peer-to-peer relationship, the weight of the edge is constructed according to the peer-to-peer duration and the peer-to-peer number, a more reasonable node sequence is generated by using a random walk strategy, vector representation of nodes is obtained according to the node sequence, association between mobile objects without the same row can be established, and similar mobile objects are further found on the basis of the peer-to-peer relationship.
Fig. 4 is a block diagram of an embodiment of a mobile object similarity determining apparatus provided in the present application, and as shown in fig. 4, the mobile object similarity determining apparatus of the present embodiment includes:
a determining module 401, configured to determine, according to trajectory data of a plurality of first moving objects, a plurality of second moving objects having a same-row relationship in the plurality of first moving objects;
a processing module 402, configured to establish a relationship graph of the plurality of second moving objects according to the plurality of second moving objects having a peer relationship; the relationship graph comprises a plurality of nodes and a plurality of edges; the node represents the second mobile object, and the edge represents that two connected nodes have a same-row relationship;
the processing module 402 is further configured to determine a node sequence using each second mobile object as an initial node according to a connection relationship between the second mobile objects in the relationship graph;
determining the vector characteristics of each second moving object by using a preset model according to a node sequence taking each second moving object as an initial node;
the processing module 402 is further configured to determine similarity of each second moving object according to the vector feature of each second moving object.
In one possible implementation, the relationship diagram further includes: a weight of the edge; the weight of the edge is determined according to the peer-to-peer duration and the peer-to-peer times of two nodes connected by the edge; the number of times of the same line is the number of times of the same line of the two nodes in a first preset time length;
the processing module 402 is specifically configured to: determining a node sequence taking each second moving object as an initial node by using a second preset model according to the connection relation among the second moving objects in the relation graph and the weight corresponding to the connection relation; the sequence of nodes includes at least one node;
the second preset model is obtained by training according to a plurality of sample data, and each sample data comprises: the relationship graph and the node sequence taking each node as a starting node in the relationship graph.
In a possible implementation manner, the processing module 402 is specifically configured to:
clustering the track data according to a preset time window to obtain at least one cluster; the class cluster comprises a plurality of first moving objects; the trajectory data includes: the position information and the time information of the first moving object at a plurality of track points;
determining a plurality of second moving objects with the same row relation from a plurality of class clusters in a second preset time length; the second preset time length is greater than or equal to the time lengths of at least two preset time windows.
In a possible implementation manner, the processing module 402 is specifically configured to:
if a plurality of second moving objects exist in the cluster of a first number of preset time windows in the second preset duration, and at least a second number of continuous preset time windows exist in the first number of preset time windows, determining that the plurality of second moving objects have a same-row relationship.
In a possible implementation manner, the number of the second moving objects having the same row relationship is greater than or equal to a third number.
In a possible implementation manner, the processing module 402 is further configured to:
preprocessing the track data to obtain preprocessed track data; the pre-treatment comprises at least one of: interpolation processing and filtering processing.
In a possible implementation manner, the processing module 402 is further configured to:
for any of the second moving objects, a fourth number of second moving objects similar to the second moving object trajectory is determined.
The apparatus of this embodiment may be configured to implement the technical solutions of the above method embodiments, and the implementation principles and technical effects are similar, which are not described herein again.
Fig. 5 is a block diagram of an embodiment of an electronic device provided in the present application, and as shown in fig. 5, the electronic device includes:
a processor 501, and a memory 502 for storing executable instructions for the processor 501.
Optionally, the method may further include: a communication interface 503 for enabling communication with other devices.
The above components may communicate over one or more buses.
The processor 501 is configured to execute the corresponding method in the foregoing method embodiment by executing the executable instruction, and the specific implementation process of the method may refer to the foregoing method embodiment, which is not described herein again.
The embodiment of the present application further provides a computer-readable storage medium, where a computer program is stored, and when the computer program is executed by a processor, the method in the foregoing method embodiment is implemented.
Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This application is intended to cover any variations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.
It will be understood that the present disclosure is not limited to the precise arrangements described above and shown in the drawings and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.
Claims (10)
1. A method for similarity determination of moving objects, comprising:
determining a plurality of second moving objects with a same-row relationship in the plurality of first moving objects according to the track data of the plurality of first moving objects;
establishing a relationship graph of the plurality of second moving objects according to the plurality of second moving objects with the same-row relationship; the relationship graph comprises a plurality of nodes and a plurality of edges; the node represents the second mobile object, and the edge represents that two connected nodes have a same-row relationship;
determining a node sequence taking each second mobile object as an initial node according to the connection relation among the second mobile objects in the relation graph;
determining the vector characteristics of each second moving object by using a first preset model according to a node sequence taking each second moving object as an initial node;
and determining the similarity of the second moving objects according to the vector characteristics of the second moving objects.
2. The method of claim 1, wherein the relationship graph further comprises: a weight of the edge; the weight of the edge is determined according to the peer-to-peer duration and the peer-to-peer times of two nodes connected by the edge; the number of times of the same line is the number of times of the same line of the two nodes in a first preset time length;
determining a node sequence taking each second mobile object as a starting node according to the connection relationship among the second mobile objects in the relationship graph, wherein the node sequence comprises:
determining a node sequence taking each second moving object as an initial node by using a second preset model according to the connection relation among the second moving objects in the relation graph and the weight corresponding to the connection relation; the sequence of nodes includes at least one node;
the second preset model is obtained by training according to a plurality of sample data, and each sample data comprises: the relationship graph and the node sequence taking each node as a starting node in the relationship graph.
3. The method according to claim 1 or 2, wherein the determining a plurality of second moving objects having a same-row relationship among a plurality of first moving objects according to trajectory data of the plurality of first moving objects comprises:
clustering the track data according to a preset time window to obtain at least one cluster; the class cluster comprises a plurality of first moving objects; the trajectory data includes: the position information and the time information of the first moving object at a plurality of track points;
determining a plurality of second moving objects with the same row relation from a plurality of class clusters in a second preset time length; the second preset time length is greater than or equal to the time lengths of at least two preset time windows.
4. The method according to claim 3, wherein the determining a plurality of second moving objects having a same-row relationship from a plurality of the class clusters within a second preset time period comprises:
if a plurality of second moving objects exist in the cluster of a first number of preset time windows in the second preset duration, and at least a second number of continuous preset time windows exist in the first number of preset time windows, determining that the plurality of second moving objects have a same-row relationship.
5. The method of claim 4,
the number of the second moving objects having the same-row relationship is greater than or equal to a third number.
6. The method of claim 3, wherein before clustering the trajectory data according to the preset time window, further comprising:
preprocessing the track data to obtain preprocessed track data; the pre-treatment comprises at least one of: interpolation processing and filtering processing.
7. The method according to claim 1 or 2, wherein after determining the similarity of each of the second moving objects, further comprising:
for any of the second moving objects, a fourth number of second moving objects similar to the second moving object trajectory is determined.
8. An apparatus for determining similarity of moving objects, comprising:
the determining module is used for determining a plurality of second moving objects with the same row relation in the plurality of first moving objects according to the track data of the plurality of first moving objects;
the processing module is used for establishing a relationship graph of the plurality of second moving objects according to the plurality of second moving objects with the same-row relationship; the relationship graph comprises a plurality of nodes and a plurality of edges; the node represents the second mobile object, and the edge represents that two connected nodes have a same-row relationship;
the processing module is further configured to determine a node sequence using each second mobile object as an initial node according to a connection relationship between each second mobile object in the relationship graph;
determining the vector characteristics of each second moving object by using a first preset model according to a node sequence taking each second moving object as an initial node;
the processing module is further configured to determine similarity of each second moving object according to the vector feature of each second moving object.
9. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the method of any one of claims 1 to 7.
10. An electronic device, comprising:
a processor; and
a memory for storing executable instructions of the processor;
wherein the processor is configured to perform the method of any of claims 1-7 via execution of the executable instructions.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010922759.0A CN112015956A (en) | 2020-09-04 | 2020-09-04 | Similarity determination method, device, equipment and storage medium for mobile object |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010922759.0A CN112015956A (en) | 2020-09-04 | 2020-09-04 | Similarity determination method, device, equipment and storage medium for mobile object |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112015956A true CN112015956A (en) | 2020-12-01 |
Family
ID=73516240
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010922759.0A Pending CN112015956A (en) | 2020-09-04 | 2020-09-04 | Similarity determination method, device, equipment and storage medium for mobile object |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112015956A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113158415A (en) * | 2021-02-23 | 2021-07-23 | 电子科技大学长三角研究院(衢州) | Vehicle track similarity evaluation method based on error analysis |
| CN116602664A (en) * | 2023-07-17 | 2023-08-18 | 青岛市胶州中心医院 | Comprehensive diagnosis and treatment nursing system for neurosurgery patients |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070237419A1 (en) * | 2006-04-11 | 2007-10-11 | Eli Shechtman | Space-time behavior based correlation |
| US20120197909A1 (en) * | 2009-10-12 | 2012-08-02 | Beel Joeran | Method for determining a similarity of objects |
| CN106383868A (en) * | 2016-09-05 | 2017-02-08 | 电子科技大学 | Road network-based spatio-temporal trajectory clustering method |
| CN106921940A (en) * | 2015-12-28 | 2017-07-04 | 宁波普天通信技术有限公司 | With track user's lookup method and device and go together user's lookup method and device |
| CN108564788A (en) * | 2018-06-07 | 2018-09-21 | 重庆邮电大学 | A kind of colleague's vehicle discovery method based on streaming big data |
| CN108595539A (en) * | 2018-04-04 | 2018-09-28 | 烟台海颐软件股份有限公司 | A kind of recognition methods of trace analogical object and system based on big data |
| CN109766786A (en) * | 2018-12-21 | 2019-05-17 | 深圳云天励飞技术有限公司 | Character relation analysis method and Related product |
| CN110348347A (en) * | 2019-06-28 | 2019-10-18 | 深圳市商汤科技有限公司 | A kind of information processing method and device, storage medium |
| CN110889422A (en) * | 2018-09-10 | 2020-03-17 | 百度在线网络技术（北京）有限公司 | Method, device and equipment for judging vehicles in same driving and computer readable medium |
| CN111143496A (en) * | 2019-12-19 | 2020-05-12 | 北京百度网讯科技有限公司 | Method and device for determining target objects with similar tracks |
| CN111191147A (en) * | 2020-04-09 | 2020-05-22 | 南京柏跃软件有限公司 | Big data mining-based group discovery algorithm model and analysis module |
| CN111612675A (en) * | 2020-05-18 | 2020-09-01 | 浙江宇视科技有限公司 | Method, device and equipment for determining peer objects and storage medium |
-
2020
- 2020-09-04 CN CN202010922759.0A patent/CN112015956A/en active Pending
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070237419A1 (en) * | 2006-04-11 | 2007-10-11 | Eli Shechtman | Space-time behavior based correlation |
| US20120197909A1 (en) * | 2009-10-12 | 2012-08-02 | Beel Joeran | Method for determining a similarity of objects |
| CN106921940A (en) * | 2015-12-28 | 2017-07-04 | 宁波普天通信技术有限公司 | With track user's lookup method and device and go together user's lookup method and device |
| CN106383868A (en) * | 2016-09-05 | 2017-02-08 | 电子科技大学 | Road network-based spatio-temporal trajectory clustering method |
| CN108595539A (en) * | 2018-04-04 | 2018-09-28 | 烟台海颐软件股份有限公司 | A kind of recognition methods of trace analogical object and system based on big data |
| CN108564788A (en) * | 2018-06-07 | 2018-09-21 | 重庆邮电大学 | A kind of colleague's vehicle discovery method based on streaming big data |
| CN110889422A (en) * | 2018-09-10 | 2020-03-17 | 百度在线网络技术（北京）有限公司 | Method, device and equipment for judging vehicles in same driving and computer readable medium |
| CN109766786A (en) * | 2018-12-21 | 2019-05-17 | 深圳云天励飞技术有限公司 | Character relation analysis method and Related product |
| CN110348347A (en) * | 2019-06-28 | 2019-10-18 | 深圳市商汤科技有限公司 | A kind of information processing method and device, storage medium |
| CN111143496A (en) * | 2019-12-19 | 2020-05-12 | 北京百度网讯科技有限公司 | Method and device for determining target objects with similar tracks |
| CN111191147A (en) * | 2020-04-09 | 2020-05-22 | 南京柏跃软件有限公司 | Big data mining-based group discovery algorithm model and analysis module |
| CN111612675A (en) * | 2020-05-18 | 2020-09-01 | 浙江宇视科技有限公司 | Method, device and equipment for determining peer objects and storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 潘晓;马昂;郭景峰;吴雷;刘风阳;: "基于时间序列的轨迹数据相似性度量方法研究及应用综述", 燕山大学学报, no. 06, pages 65 - 79 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113158415A (en) * | 2021-02-23 | 2021-07-23 | 电子科技大学长三角研究院(衢州) | Vehicle track similarity evaluation method based on error analysis |
| CN113158415B (en) * | 2021-02-23 | 2023-09-08 | 电子科技大学长三角研究院(衢州) | Vehicle track similarity evaluation method based on error analysis |
| CN116602664A (en) * | 2023-07-17 | 2023-08-18 | 青岛市胶州中心医院 | Comprehensive diagnosis and treatment nursing system for neurosurgery patients |
| CN116602664B (en) * | 2023-07-17 | 2023-09-22 | 青岛市胶州中心医院 | Comprehensive diagnosis and treatment nursing system for neurosurgery patients |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110476176B (en) | User objective assistance techniques | |
| CN110634291B (en) | High-precision map topology automatic construction method and system based on crowdsourcing data | |
| Liu et al. | Cdas: a crowdsourcing data analytics system | |
| Gupte et al. | A guide to pre‐processing high‐throughput animal tracking data | |
| CN109446302A (en) | Question and answer data processing method, device and computer equipment based on machine learning | |
| US20120150871A1 (en) | Autonomous Mobile Blogging | |
| Kacorri et al. | Supporting orientation of people with visual impairment: Analysis of large scale usage data | |
| CN109409971A (en) | Abnormal order processing method and device | |
| CN112015956A (en) | Similarity determination method, device, equipment and storage medium for mobile object | |
| CN110472154A (en) | A kind of resource supplying method, apparatus, electronic equipment and readable storage medium storing program for executing | |
| CN111915020A (en) | Method and device for updating detection model and storage medium | |
| CN112815948B (en) | Method, device, computer equipment and storage medium for identifying yaw mode | |
| CN110610182A (en) | User track similarity judgment method and related device | |
| CN113587944B (en) | Quasi-real-time vehicle driving route generation method, system and equipment | |
| Atzmueller | Detecting community patterns capturing exceptional link trails | |
| CN110208001A (en) | A kind of road test method and apparatus of vehicle | |
| CN112885099A (en) | Method, device and equipment for determining key points of vehicle track | |
| CN111258974A (en) | Vehicle offline scene data processing method and system | |
| CN113656686A (en) | Task report generation method based on birth teaching fusion and service system | |
| CN113689291B (en) | Anti-fraud identification method and system based on abnormal movement | |
| CN113327103B (en) | Intelligent campus epidemic situation on-line monitoring and early warning method, system and storage medium | |
| US20180285758A1 (en) | Methods for creating and analyzing dynamic trail networks | |
| CN113761355A (en) | Information recommendation method, device, equipment and computer readable storage medium | |
| CN111325255B (en) | Specific crowd delineating method and device, electronic equipment and storage medium | |
| CN114394099B (en) | Method and device for identifying abnormal running of vehicle, computer equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |