CN107862862B - Vehicle behavior analysis method and device - Google Patents

Abstract

The embodiment of the invention discloses a vehicle behavior analysis method and a device, wherein the method comprises the following steps: obtaining a vehicle track of a vehicle to be analyzed in first preset time, wherein the vehicle track comprises a bayonet mark of a bayonet through which the vehicle to be analyzed passes; classifying the vehicle track of the vehicle to be analyzed according to a preset theme model; and for each type of vehicle track, determining behavior information of the vehicle to be analyzed in first preset time according to the gate information corresponding to the gate identification contained in each vehicle track. Therefore, according to the preset theme model, the intrinsic relation of the vehicle track of the vehicle to be analyzed in the first preset time is determined, and further, according to the intrinsic relation of the vehicle track, the behavior information of the vehicle to be analyzed in the first preset time is determined. And subsequently, the vehicle to be analyzed is tracked according to the behavior information, so that the accuracy of tracking the vehicle to be analyzed can be improved.

Description

Vehicle behavior analysis method and device

Technical Field

The invention relates to the field of data mining, in particular to a vehicle behavior analysis method and device.

Background

In the modern times, vehicles are becoming more and more the tools of people for riding instead of walk. The large amount of vehicle information (such as vehicle passing records and the like) collected by the monitoring devices on the road is generally only used for searching for the illegal vehicle, and the implicit information in the large amount of vehicle information collected by each monitoring device is not concerned. Such as: determining the vehicle trajectory of the vehicle based on the large amount of vehicle information collected by the monitoring device, and the like.

As the vehicle is more and more a transportation tool of people in the current times, the vehicle track of the vehicle is further analyzed, the behavior information of the vehicle can be further determined, and some relevant information of a driver, such as the information of interests, hobbies, behavior habits and the like of people, can be further determined.

Then, how to analyze the behavior of the vehicle for a large amount of vehicle information collected by the monitoring device on the road becomes an urgent problem to be solved.

Disclosure of Invention

The embodiment of the invention discloses a vehicle behavior analysis method and device, which aim to obtain behavior information of a corresponding vehicle according to a vehicle track of the vehicle. The specific scheme is as follows:

in one aspect, an embodiment of the present invention provides a vehicle behavior analysis method, where the method includes:

obtaining a vehicle track of a vehicle to be analyzed in first preset time, wherein the vehicle track comprises a bayonet identification of a bayonet through which the vehicle to be analyzed passes;

classifying the vehicle track of the vehicle to be analyzed according to a preset theme model;

and for each type of vehicle track, determining behavior information of the vehicle to be analyzed in the first preset time according to the checkpoint information corresponding to the checkpoint identification contained in each vehicle track.

Optionally, the obtaining a vehicle trajectory of the vehicle to be analyzed within the first predetermined time includes:

obtaining first vehicle passing record data of the vehicle to be analyzed passing through a gate in first preset time;

and determining the vehicle track of the vehicle to be analyzed according to a preset track division rule and the first vehicle passing record data.

Optionally, the classifying the vehicle trajectory of the vehicle to be analyzed according to the preset topic model includes:

inputting the vehicle track of the vehicle to be analyzed into the theme model;

obtaining a theme corresponding to each vehicle track output by the theme model;

the vehicle trajectories with the same theme are classified into one category.

Optionally, the topic model is a latent dirichlet allocation LDA model; the LDA model comprises a sample bayonet identification, a theme and a sample vehicle track three-layer structure of the sample bayonet;

the method further includes a process of building a topic model, the process including:

obtaining sample vehicle-passing record data of each sample vehicle of each sample gate in a second preset time, wherein the sample vehicle-passing record data comprises sample identification information of the sample vehicle;

determining sample vehicle-passing record data corresponding to the sample identification information of each sample vehicle from the sample vehicle-passing record data;

determining a sample vehicle track of a corresponding sample vehicle according to a preset track division rule and sample vehicle-passing record data corresponding to the same sample identification information, wherein the sample vehicle track comprises a sample gate identification of a sample gate through which the vehicle passes;

performing LDA learning to obtain the corresponding relation between each sample gate mark and the theme and the conversion relation between the theme of the sample gate mark and the theme of the sample vehicle track;

the obtaining of the theme corresponding to each vehicle track output by the theme model includes:

the theme model determines a theme corresponding to the bayonet identification according to the corresponding relation between each sample bayonet identification and the theme aiming at the bayonet identification of each input vehicle track, determines the theme of each vehicle track according to the conversion relation between the theme of the sample bayonet identification and the theme of the sample vehicle track, and outputs the determined theme of each vehicle track;

a theme of each of the outputted vehicle trajectories is obtained.

Optionally, the first vehicle passing record data includes a first vehicle passing time when the vehicle to be analyzed passes through the corresponding gate;

the determining the vehicle track of the vehicle to be analyzed according to the preset track division rule and the first vehicle passing record data comprises the following steps:

sequencing the first vehicle passing record data according to the corresponding first vehicle passing time sequence;

determining a first time difference of every two adjacent first vehicle passing record data according to the first vehicle passing time corresponding to each first vehicle passing record data;

when the first time difference exceeds a preset first track division time threshold value, determining two corresponding adjacent first vehicle passing record data as a first track division limit, wherein a bayonet mark of a bayonet corresponding to the first vehicle passing record data before the first vehicle passing time is a previous vehicle track end mark, and a bayonet mark of a bayonet corresponding to the first vehicle passing record data after the first vehicle passing time is a next vehicle track start mark;

and determining the vehicle track of the vehicle to be analyzed according to the first track dividing boundary and the first vehicle passing record data.

Optionally, the sample vehicle-passing record data further includes a second vehicle-passing time when the corresponding sample vehicle passes through the corresponding gate;

the determining the sample vehicle track of the corresponding sample vehicle according to the preset track division rule and the sample vehicle-passing record data corresponding to the same sample identification information includes:

sequencing the sample vehicle-passing record data according to the corresponding second vehicle-passing time sequence;

determining a second time difference of every two adjacent sample vehicle-passing record data according to a second vehicle-passing time corresponding to each sample vehicle-passing record data;

when the second time difference exceeds a preset second track division time threshold value, taking the corresponding two adjacent sample vehicle-passing record data as a second track division boundary, wherein a bayonet mark of a bayonet corresponding to the vehicle-passing record data before the second vehicle-passing time is a previous vehicle track end mark, and a bayonet mark of a bayonet corresponding to the vehicle-passing record data after the second vehicle-passing time is a next vehicle track start mark;

and determining the vehicle track of the corresponding sample vehicle according to the second track dividing boundary and the sample vehicle passing record data.

Optionally, the card port information includes: longitude and latitude information of the bayonet corresponding to each bayonet identification, building information in a preset range of the bayonet corresponding to each bayonet identification or environment information of the bayonet corresponding to each bayonet identification;

for each type of vehicle track, determining behavior information of the vehicle to be analyzed in the first preset time according to the gate information corresponding to the gate identifier contained in each vehicle track, including:

determining a region theme corresponding to each type of vehicle track according to longitude and latitude information corresponding to the gate identification contained in each vehicle track;

determining a region theme corresponding to the vehicle to be analyzed within the first preset time according to the region theme corresponding to each determined vehicle track; or the like, or, alternatively,

determining a building theme corresponding to each type of vehicle track according to building information in a bayonet preset range corresponding to a bayonet mark contained in each vehicle track;

determining a building theme corresponding to the vehicle to be analyzed within the first preset time according to the building theme corresponding to each determined vehicle track; or the like, or, alternatively,

determining an environment theme corresponding to each type of vehicle track according to environment information of a bayonet corresponding to a bayonet identification contained in each vehicle track;

and determining an environment theme corresponding to the vehicle to be analyzed within the first preset time according to the environment theme corresponding to each determined vehicle track.

In another aspect, an embodiment of the present invention provides a vehicle behavior analysis apparatus, including:

the analysis method comprises the steps of obtaining a vehicle track of a vehicle to be analyzed in first preset time, wherein the vehicle track comprises a bayonet identification of a bayonet through which the vehicle to be analyzed passes;

the classification module is used for classifying the vehicle track of the vehicle to be analyzed according to a preset theme model;

and the determining module is used for determining the behavior information of the vehicle to be analyzed in the first preset time according to the vehicle track of each type and the vehicle track information corresponding to the vehicle mount identifier contained in each vehicle track.

Optionally, the obtaining module includes a first obtaining unit and a first determining unit;

the first obtaining unit is used for obtaining first vehicle passing record data of the vehicle to be analyzed passing a gate within first preset time, wherein the first vehicle passing record data comprises identification information of the corresponding vehicle;

the first determining unit is used for determining the vehicle track of the vehicle to be analyzed according to a preset track division rule and the first vehicle passing record data.

Optionally, the classification module includes an input unit, a second obtaining unit, and a dividing unit;

the input unit is used for inputting the vehicle track of the vehicle to be analyzed into the theme model;

the second obtaining unit is used for obtaining a theme corresponding to each vehicle track output by the theme model;

the dividing unit is used for dividing the vehicle tracks with the same theme into one type.

Optionally, the topic model is a latent dirichlet allocation LDA model; the LDA model comprises a sample bayonet identification, a theme and a sample vehicle track three-layer structure of the sample bayonet;

the device also comprises a theme model establishing module used for establishing a theme model, wherein the theme model establishing module comprises a third obtaining unit, a second determining unit, a third determining unit and a fourth obtaining unit;

the third obtaining unit is used for obtaining sample vehicle-passing record data of each sample vehicle of each sample gate in a second preset time, wherein the sample vehicle-passing record data comprises sample identification information of the sample vehicle;

the second determining unit is used for determining sample vehicle-passing record data corresponding to the sample identification information of each sample vehicle from the sample vehicle-passing record data;

the third determining unit is used for determining a sample vehicle track of the corresponding sample vehicle according to a preset track division rule and sample vehicle-passing record data corresponding to the same sample identification information, wherein the sample vehicle track comprises a sample gate identification of a sample gate through which the vehicle passes;

the fourth obtaining unit is used for performing LDA learning to obtain a corresponding relation between each sample gate identifier and the theme and a conversion relation between the theme of the sample gate identifier and the theme of the sample vehicle track;

the second obtaining unit is specifically configured to, for the bayonet identifier of each input vehicle track, determine a theme corresponding to the bayonet identifier according to a correspondence between each sample bayonet identifier and the theme, determine a theme of each vehicle track according to a conversion relationship between the theme of the sample bayonet identifier and the theme of the sample vehicle track, and output the determined theme of each vehicle track;

a theme of each of the outputted vehicle trajectories is obtained.

Optionally, the first vehicle passing record data includes a first vehicle passing time when the vehicle to be analyzed passes through the corresponding gate;

the first determining unit comprises a first sequencing submodule, a first determining submodule, a second determining submodule and a third determining submodule;

the first sequencing submodule is used for sequencing the first vehicle passing record data according to the corresponding first vehicle passing time sequence;

the first determining submodule is used for determining a first time difference between every two adjacent first vehicle passing record data according to the first vehicle passing time corresponding to each first vehicle passing record data;

the second determining submodule is used for determining two corresponding adjacent first vehicle passing record data as a first track dividing boundary when the first time difference exceeds a preset first track dividing time threshold, wherein a bayonet mark of a bayonet corresponding to the first vehicle passing record data before the first vehicle passing time is a previous vehicle track end mark, and a bayonet mark of a bayonet corresponding to the first vehicle passing record data after the first vehicle passing time is a next vehicle track start mark;

the third determining submodule is used for determining the vehicle track of the vehicle to be analyzed according to the first track dividing boundary and the first vehicle passing record data.

Optionally, the sample vehicle-passing record data further includes a second vehicle-passing time when the corresponding sample vehicle passes through the corresponding gate;

the third determining unit comprises a second sequencing submodule, a fourth determining submodule, a fifth determining submodule and a sixth determining submodule;

the second sequencing submodule is used for sequencing the sample vehicle-passing record data according to the corresponding second vehicle-passing time sequence;

the fourth determining submodule is used for determining a second time difference of every two adjacent sample vehicle-passing record data according to a second vehicle-passing time corresponding to each sample vehicle-passing record data;

the fifth determining submodule is used for taking the corresponding two adjacent sample vehicle-passing record data as a second track dividing boundary when the second time difference exceeds a preset second track dividing time threshold, wherein a bayonet mark of a bayonet corresponding to the vehicle-passing record data before the second vehicle-passing time is a previous vehicle track end mark, and a bayonet mark of a bayonet corresponding to the vehicle-passing record data after the second vehicle-passing time is a next vehicle track start mark;

and the sixth determining submodule is used for determining the vehicle track of the corresponding sample vehicle according to the second track dividing boundary and the sample vehicle passing record data.

Optionally, the card port information includes: longitude and latitude information of the bayonet corresponding to each bayonet identification, building information in a preset range of the bayonet corresponding to each bayonet identification or environment information of the bayonet corresponding to each bayonet identification;

the determining module is specifically used for determining a region theme corresponding to each type of vehicle track according to longitude and latitude information corresponding to the gate identification contained in each vehicle track;

determining a region theme corresponding to the vehicle to be analyzed within the first preset time according to the region theme corresponding to each determined vehicle track; or the like, or, alternatively,

determining a building theme corresponding to each type of vehicle track according to building information in a bayonet preset range corresponding to a bayonet mark contained in each vehicle track;

determining a building theme corresponding to the vehicle to be analyzed within the first preset time according to the building theme corresponding to each determined vehicle track; or the like, or, alternatively,

determining an environment theme corresponding to each type of vehicle track according to environment information of a bayonet corresponding to a bayonet identification contained in each vehicle track;

and determining an environment theme corresponding to the vehicle to be analyzed within the first preset time according to the environment theme corresponding to each determined vehicle track.

In the scheme, a vehicle track of a vehicle to be analyzed in a first preset time is obtained, wherein the vehicle track comprises a bayonet identification of a bayonet through which the vehicle to be analyzed passes; classifying the vehicle track of the vehicle to be analyzed according to a preset theme model; and for each type of vehicle track, determining behavior information of the vehicle to be analyzed in first preset time according to the gate information corresponding to the gate identification contained in each vehicle track. Therefore, according to the preset theme model, the intrinsic relation of the vehicle track of the vehicle to be analyzed in the first preset time is determined, and further, according to the intrinsic relation of the vehicle track, the behavior information of the vehicle to be analyzed in the first preset time is determined. The vehicle to be analyzed can be tracked subsequently according to the behavior information, and the accuracy of tracking the vehicle to be analyzed can be improved. Of course, it is not necessary for any product or method of practicing the invention to achieve all of the above-described advantages at the same time.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.

Fig. 1 is a schematic flow chart of a vehicle behavior analysis method according to an embodiment of the present invention;

FIG. 2A is a flowchart illustrating a vehicle trajectory classification process according to an embodiment of the present invention;

FIG. 2B is an exemplary illustration of vehicle trajectory segmentation in an embodiment of the present invention;

FIG. 3 is a schematic flow chart of topic model creation in an embodiment of the present invention;

fig. 4 is a schematic structural diagram of a vehicle behavior analysis apparatus according to an embodiment of the present invention;

FIG. 5 is a schematic structural diagram of a topic model building module in an embodiment of the present invention;

fig. 6 is a diagram illustrating an exemplary configuration of a vehicle trajectory segmentation module according to an embodiment of the present invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

The embodiment of the invention provides a vehicle behavior analysis method and device, which are used for obtaining behavior information of a corresponding vehicle according to a vehicle track of the vehicle.

First, a vehicle behavior analysis method according to an embodiment of the present invention will be described.

As shown in fig. 1, an embodiment of the present invention provides a vehicle behavior analysis method, which may include the following steps:

s101: obtaining a vehicle track of a vehicle to be analyzed in first preset time, wherein the vehicle track comprises a bayonet identification of a bayonet through which the vehicle to be analyzed passes;

it is understood that each vehicle track of the vehicle to be analyzed may include at least two bayonet identifications, and the first predetermined time may be half a year, or the like, and the first predetermined time is not limited in the embodiments of the present invention. The vehicle trajectory may be determined in various ways, such as determining the vehicle trajectory according to the positioning information of the corresponding vehicle.

In a specific implementation manner, determining a vehicle trajectory of the vehicle to be analyzed according to a time sequence of passing gates of the vehicle to be analyzed and gate identifications of the passing gates, where the obtaining a vehicle trajectory of the vehicle to be analyzed within a first predetermined time (S101) may include:

obtaining first vehicle passing record data of a vehicle to be analyzed passing a gate within first preset time;

and determining the vehicle track of the vehicle to be analyzed according to a preset track division rule and the first vehicle passing record data.

It should be noted that each bayonet is provided with a corresponding video monitoring device or scanning device, and the video monitoring device or scanning device can record vehicle passing record data of each vehicle passing through the corresponding bayonet, wherein the vehicle passing record data includes identification information capable of distinguishing different vehicles. The obtaining of the first vehicle passing record data may adopt the prior art, which is not described herein.

In a specific implementation manner, the first vehicle passing record data includes a first vehicle passing time when the vehicle to be analyzed passes through a corresponding gate;

as shown in fig. 2A, the determining the vehicle trajectory of the vehicle to be analyzed according to the preset trajectory partition rule and the first vehicle passing record data may include:

s201: sequencing the first vehicle passing record data according to the corresponding first vehicle passing time sequence;

s202: determining a first time difference of every two adjacent first vehicle passing record data according to the first vehicle passing time corresponding to each first vehicle passing record data;

s203: when the first time difference exceeds a preset first track division time threshold value, determining two corresponding adjacent first vehicle passing record data as a first track division limit, wherein a bayonet mark of a bayonet corresponding to the first vehicle passing record data before the first vehicle passing time is a previous vehicle track end mark, and a bayonet mark of a bayonet corresponding to the first vehicle passing record data after the first vehicle passing time is a next vehicle track start mark;

s204: and determining the vehicle track of the vehicle to be analyzed according to the first track dividing boundary and the first vehicle passing record data.

In addition to the identification information of the corresponding vehicle, the vehicle passing record data of the vehicle recorded by the video monitoring device corresponding to each gate may also include a vehicle passing time of the corresponding vehicle, specifically, the first vehicle passing record data includes a first vehicle passing time when the vehicle to be analyzed passes through the corresponding gate.

Through the first passing time, the sequence of the vehicle to be analyzed passing through each bayonet can be determined, and then the vehicle track of the vehicle to be analyzed can be determined. It can be understood that the vehicle track is divided according to the first vehicle passing record data of the vehicle to be analyzed passing through each gate, and the division can be determined according to the staying time of the vehicle to be analyzed in a certain place. Generally, if the vehicle is always in a driving state, the time interval between every two passing gates is generally about several minutes to several tens of minutes. In other words, when the time interval for the vehicle to pass through the two gates is relatively large (e.g., more than two hours), it may be determined that the vehicle has made a stop between the two gates, and the larger the time interval for the vehicle to pass through the two gates may indicate that the vehicle has made a longer stop between the two gates. At this time, the two corresponding bayonets may be used as the first trajectory partition boundary, the bayonet identifier corresponding to the first vehicle passing record data before the first vehicle passing time is the previous vehicle trajectory end identifier, and the bayonet identifier corresponding to the first vehicle passing record data after the first vehicle passing time is the next vehicle trajectory start identifier.

For example, as shown in fig. 2B, an exemplary diagram of the vehicle trajectory division provided by the embodiment of the invention is shown; the coordinates shown in fig. 2B are time scales of a day, and first vehicle passing record data of a passing gate of the vehicle a to be analyzed in the day is marked on the coordinates, where a gate identifier corresponding to the gate and identification information of the vehicle a to be analyzed are recorded in the first vehicle passing record data, and the method specifically includes: a bayonet sign of 19 minutes-bayonet 1 at a-6 time, specifically a bayonet sign that a vehicle to be analyzed A6 passes through bayonet 1 at 19 minutes (first passing time), and 1 minutes-bayonet 2 at a-7 time, specifically a bayonet sign that a vehicle to be analyzed A7 passes through bayonet 2 at 1 minutes (first passing time), and 50 minutes-bayonet 3 at a-7 time, specifically a bayonet sign that a vehicle to be analyzed A7 passes through bayonet 3 at 50 minutes (first passing time), and a bayonet sign that a vehicle to be analyzed A11 passes through bayonet 4 at 50 minutes (first passing time), specifically a bayonet sign that a vehicle to be analyzed A11 passes through bayonet 4 at 50 minutes (first passing time), a bayonet sign that a vehicle to be analyzed A12 passes through bayonet 5 at 40 minutes (first passing time), specifically a bayonet sign that a vehicle to be analyzed A12 passes through bayonet 5 at 40 minutes (first passing time), and a bayonet sign that a vehicle to 13 passes through bayonet 6 at 28 minutes, the bayonet identification specifically representing that a vehicle A13 to be analyzed passes through a bayonet 6 in 28 minutes (first passing time) and a bayonet 3 in 20 minutes-A-17, specifically representing that a vehicle A17 to be analyzed passes through a bayonet 3 in 20 minutes (first passing time) and a bayonet 2 in 58 minutes-A-17, specifically representing that a vehicle A17 to be analyzed passes through a bayonet 2 in 58 minutes (first passing time) and a bayonet 1 in 52 minutes-A-18, specifically representing that a vehicle A18 to be analyzed passes through a bayonet 1 in 52 minutes (first passing time); wherein the preset first track division time threshold is 2 hours.

And respectively calculating first time difference between every two adjacent gates corresponding to the first passing time of the vehicle A to be analyzed. Each calculated first time difference is compared to a first trajectory split time threshold (2 hours). Wherein it can be determined that the first time difference between passing bayonet 3 and passing bayonet 4 exceeds the first trajectory division time threshold (2 hours), which is (7 hours, 50 minutes and 11 hours, 50 minutes) for 4 hours; the first time difference between the passing bayonet 6 and the passing bayonet 3 exceeds a first track division time threshold value (2 hours), and is (20 minutes at 17 hours to 13 minutes and 28 minutes at 28 hours) for 3 hours and 52 minutes; and respectively determining the corresponding two adjacent first vehicle passing record data as first track division boundaries.

Specifically, the three vehicle tracks can be divided into 1-2-3 according to the first vehicle passing record data; 4-5-6; 3-2-1.

S102: classifying the vehicle track of the vehicle to be analyzed according to a preset theme model;

it can be understood that, the subject model has been learned by training a large number of bayonet markers and their corresponding categories, and vehicle trajectories including the bayonet markers and their corresponding categories, and the vehicle trajectories of the vehicle to be analyzed can be classified according to the training results learned by the subject model. Wherein, the topic model may adopt the LDA (latent dirichletaillocation, implicit dirichlet distribution) model of the prior art.

In a specific implementation manner, the classifying the vehicle trajectory of the vehicle to be analyzed according to the preset topic model may include:

inputting the vehicle track of the vehicle to be analyzed into the theme model;

obtaining a theme corresponding to each vehicle track output by the theme model;

the vehicle trajectories with the same theme are classified into one category.

It should be noted that the theme model may continue to train the inputted vehicle tracks of the vehicle to be analyzed according to the pre-learned training result, train the theme corresponding to each vehicle track, and divide the vehicle tracks with the same corresponding theme into one class, wherein, according to the existing theme model technology, each vehicle track may belong to different classes of themes at the same time, and the probabilities of the class of themes in which the vehicle tracks are located may be different or the same. I.e., each vehicle trajectory may correspond to one or more categories of subject matter. The theme model is used for determining the probability that the input information belongs to a certain theme based on the implicit association of the input information, if the input information is a vehicle track, the vehicle track comprises a bayonet mark, and the bayonet mark is the implicit association between the vehicle tracks. Further, the input information is clustered or classified according to the probability that the input information belongs to a certain topic.

S103: and for each type of vehicle track, determining behavior information of the vehicle to be analyzed in the first preset time according to the checkpoint information corresponding to the checkpoint identification contained in each vehicle track.

For the vehicle to be analyzed, there are multiple vehicle tracks corresponding to the vehicle to be analyzed within the first predetermined time. And determining a theme corresponding to the vehicle to be analyzed according to each type of vehicle track. The vehicle tracks corresponding to the vehicle to be analyzed may belong to different categories of subjects, and the probability that the corresponding vehicle tracks belong to a certain subject is higher or the proportion of the vehicle tracks of a certain subject to all the vehicle tracks of the vehicle to be analyzed is higher, so that the probability that the vehicle to be analyzed belongs to the certain subject is higher. Because each vehicle track is composed of at least two bayonet identifications, specific meanings can be defined for each type of theme according to bayonet information corresponding to the bayonet identification contained in each vehicle track, and further, behavior information of the vehicle to be analyzed in the first preset time can be continuously determined. It is understood that the behavior information may include a probability magnitude of the vehicle to be analyzed appearing in a certain type of place within a first predetermined time; or the amount of probability that it occurs in a region.

Further, the more the obtained vehicle tracks corresponding to the vehicle to be analyzed, the more accurate the determined behavior information of the vehicle to be analyzed is.

In a specific implementation manner, the bayonet information may include: longitude and latitude information of the bayonet corresponding to each bayonet identification, building information in a preset range of the bayonet corresponding to each bayonet identification or environment information of the bayonet corresponding to each bayonet identification;

for each type of vehicle track, determining behavior information of the vehicle to be analyzed within the first preset time according to the gate information corresponding to the gate identifier included in each vehicle track (S103), including:

determining a region theme corresponding to each type of vehicle track according to longitude and latitude information corresponding to the gate identification contained in each vehicle track;

determining a region theme corresponding to the vehicle to be analyzed within the first preset time according to the region theme corresponding to each determined vehicle track; or the like, or, alternatively,

determining a building theme corresponding to each type of vehicle track according to building information in a bayonet preset range corresponding to a bayonet mark contained in each vehicle track;

determining a building theme corresponding to the vehicle to be analyzed within the first preset time according to the building theme corresponding to each determined vehicle track; or the like, or, alternatively,

determining an environment theme corresponding to each type of vehicle track according to environment information of a bayonet corresponding to a bayonet identification contained in each vehicle track;

and determining an environment theme corresponding to the vehicle to be analyzed within the first preset time according to the environment theme corresponding to each determined vehicle track.

The method includes the steps that the prior art can be adopted, the specific meaning of the theme corresponding to the bayonet identification can be determined according to bayonet information corresponding to the bayonet identification, it can be understood that when the bayonet information corresponding to the bayonet identification changes, the specific meaning of the theme corresponding to the bayonet identification also changes, if the bayonet information is longitude and latitude information of the bayonet corresponding to the bayonet identification, the theme type output by the theme model corresponds to a region theme, the region theme can comprise a country category or a province category and the like, further, behavior information of the vehicle to be analyzed is analyzed, and the behavior information can be a frequent activity region range of the vehicle to be analyzed or a probability of the vehicle to be analyzed appearing in a certain activity region range. If the bayonet information is the building information in the bayonet predetermined range corresponding to the bayonet identification, the theme type output by the theme model is correspondingly a building theme, the building theme can comprise an ancient wind building theme, a modern building theme and the like, and further, the interest and hobbies, the activity and place range and the like corresponding to the vehicle (personnel) to be analyzed are analyzed. If the information of the gate is the environmental information of the gate corresponding to the gate identifier, the type of the theme output by the theme model may correspond to the theme such as entertainment, dining, sports, social contact, and the like, and further, the behavior information corresponding to the vehicle to be analyzed is analyzed, and the behavior information may include the probability or activity range of the vehicle to be analyzed appearing in a certain place (such as entertainment, dining, sports, social contact), and the like.

It can be understood that, subsequently, according to the behavior information of the vehicle to be analyzed, the accuracy of tracking the vehicle to be analyzed can be improved. Such as: when the theme corresponding to the vehicle to be analyzed is determined to be entertainment, the fact that the vehicle to be analyzed can appear at certain card ports related to the entertainment theme more frequently can be shown; the probability of tracking the vehicle to be analyzed at the casino will be greater. In addition, when the vehicle to be analyzed is determined to be frequently present at a certain gate, certain theme activities performed by the vehicle to be analyzed can be deduced with a high probability.

By applying the embodiment of the invention, the vehicle track of the vehicle to be analyzed in the first preset time is obtained, wherein the vehicle track comprises the bayonet identification of the bayonet through which the vehicle to be analyzed passes; classifying the vehicle track of the vehicle to be analyzed according to a preset theme model; and for each type of vehicle track, determining behavior information of the vehicle to be analyzed in first preset time according to the gate information corresponding to the gate identification contained in each vehicle track. Therefore, according to the preset theme model, the intrinsic relation of the vehicle track of the vehicle to be analyzed in the first preset time is determined, and further, according to the intrinsic relation of the vehicle track, the behavior information of the vehicle to be analyzed in the first preset time is determined. The vehicle to be analyzed can be tracked subsequently according to the behavior information, and the accuracy of tracking the vehicle to be analyzed can be improved.

In a specific implementation manner, the topic model is a latent dirichlet allocation LDA model; the LDA model comprises a sample bayonet identification, a theme and a sample vehicle track three-layer structure of the sample bayonet;

as shown in fig. 3, the vehicle behavior analysis method provided by the embodiment of the present invention further includes a process of establishing a topic model, where the process may include:

s301: obtaining sample vehicle-passing record data of each sample vehicle of each sample gate in a second preset time, wherein the sample vehicle-passing record data comprises sample identification information of the sample vehicle;

s302: determining sample vehicle-passing record data corresponding to the sample identification information of each sample vehicle from the sample vehicle-passing record data;

s303: determining a sample vehicle track of a corresponding sample vehicle according to a preset track division rule and sample vehicle-passing record data corresponding to the same sample identification information, wherein the sample vehicle track comprises a sample gate identification of a sample gate through which the vehicle passes;

s304: performing LDA learning to obtain the corresponding relation between each sample gate mark and the theme and the conversion relation between the theme of the sample gate mark and the theme of the sample vehicle track;

the obtaining of the theme corresponding to each vehicle track output by the theme model includes:

the theme model determines a theme corresponding to the bayonet identification according to the corresponding relation between each sample bayonet identification and the theme aiming at the bayonet identification of each input vehicle track, determines the theme of each vehicle track according to the conversion relation between the theme of the sample bayonet identification and the theme of the sample vehicle track, and outputs the determined theme of each vehicle track;

a theme of each of the outputted vehicle trajectories is obtained.

The LDA model comprises three-layer structures of sample bayonet identification, theme and sample vehicle track of a sample bayonet, wherein the corresponding relation can be expressed as sample bayonet identification-theme, theme-sample vehicle track and sample bayonet identification-sample vehicle track.

And obtaining sample vehicle-passing record data of each sample vehicle within second preset time, determining a sample vehicle track corresponding to each sample vehicle according to a preset track division rule and the sample vehicle-passing record data of each sample vehicle, and performing LDA learning on the sample vehicle track, namely starting to train the sample vehicle track. It can be understood that the correspondence between the sample mount mark and the sample vehicle track is determined, that is, the probability that the mount mark belongs to a certain vehicle track is determined, and the probability that a certain mount mark exists in a certain vehicle track is determined. A predetermined number of themes, such as theme 1, theme 2 … …, theme N-1, theme N, may be preset. As in the prior art, the LDA model firstly randomly assigns values to sample bayonet identification-subject and subject-sample vehicle tracks, and the assignment process is continuously repeated. And outputting convergence results corresponding to the sample bayonet identification-theme and the theme-sample vehicle track until the numerical values corresponding to the sample bayonet identification-theme and the theme-sample vehicle track converge to a certain result without changing. The output convergence result is the corresponding relation between each sample gate mark and the theme and the conversion relation between the theme of the sample gate mark and the theme of the sample vehicle track, the information is obtained, and the theme model is established.

Further, when the vehicle track of the vehicle to be analyzed is trained, the theme of each vehicle track corresponding to the vehicle to be analyzed is determined according to the output corresponding relation between each sample bayonet identification and the theme and the conversion relation between the theme of the sample bayonet identification and the theme of the sample vehicle track.

Wherein the second predetermined time may be the same as or different from the first predetermined time, or may have an overlap.

In a specific implementation manner, the sample vehicle-passing record data further includes a second vehicle-passing time when the corresponding vehicle passes through the corresponding gate;

the determining the sample vehicle track of the corresponding sample vehicle according to the preset track division rule and the sample vehicle-passing record data corresponding to the same sample identification information includes:

sequencing the sample vehicle-passing record data according to the corresponding second vehicle-passing time sequence;

determining a second time difference of every two adjacent sample vehicle-passing record data according to a second vehicle-passing time corresponding to each sample vehicle-passing record data;

when the second time difference exceeds a preset second track division time threshold value, taking the corresponding two adjacent sample vehicle-passing record data as a second track division boundary, wherein a bayonet mark of a bayonet corresponding to the vehicle-passing record data before the second vehicle-passing time is a previous vehicle track end mark, and a bayonet mark of a bayonet corresponding to the vehicle-passing record data after the second vehicle-passing time is a next vehicle track start mark;

and determining the vehicle track of the corresponding vehicle according to the second track dividing boundary and the sample vehicle passing record data.

It can be understood that the division of the vehicle trajectory for the sample vehicle is the same as the division of the vehicle trajectory for the vehicle to be analyzed, and details are not repeated here. The obtained sample vehicle-passing record data comprises identification information of corresponding sample vehicles. To avoid errors in dividing the track.

According to the embodiment of the invention, the LDA model is applied to conduct behavior analysis on the vehicle to be analyzed, the LDA model automatically trains the vehicle track corresponding to the vehicle to be analyzed according to the learned training result, the theme corresponding to the vehicle track is determined, the specific meaning of the theme is limited according to the bayonet information corresponding to the bayonet identification contained in the vehicle track, further, the behavior information of the vehicle to be analyzed is determined, a large amount of manpower is not required to be invested to construct the theme base, and the personnel time is effectively saved. And moreover, the LDA model is applied to conduct behavior analysis on the vehicle to be analyzed, interference of human factors is avoided, and accuracy of the behavior analysis on the vehicle is improved.

Corresponding to the above method embodiment, as shown in fig. 4, an embodiment of the present invention further provides a vehicle behavior analysis apparatus, which may include:

an obtaining module 410, configured to obtain a vehicle trajectory of a vehicle to be analyzed within a first predetermined time, where the vehicle trajectory includes a gate identifier of a gate through which the vehicle to be analyzed passes;

the classification module 420 is configured to classify the vehicle trajectory of the vehicle to be analyzed according to a preset topic model;

the determining module 430 is configured to determine, for each type of vehicle track, behavior information of the vehicle to be analyzed within the first predetermined time according to the gate information corresponding to the gate identifier included in each vehicle track.

By applying the embodiment of the invention, the vehicle track of the vehicle to be analyzed in the first preset time is obtained, wherein the vehicle track comprises the bayonet identification of the bayonet through which the vehicle to be analyzed passes; classifying the vehicle track of the vehicle to be analyzed according to a preset theme model; and for each type of vehicle track, determining behavior information of the vehicle to be analyzed in first preset time according to the gate information corresponding to the gate identification contained in each vehicle track. Therefore, according to the preset theme model, the intrinsic relation of the vehicle track of the vehicle to be analyzed in the first preset time is determined, and further, according to the intrinsic relation of the vehicle track, the behavior information of the vehicle to be analyzed in the first preset time is determined. The vehicle to be analyzed can be tracked subsequently according to the behavior information, and the accuracy of tracking the vehicle to be analyzed can be improved.

In a specific implementation manner, the obtaining module 410 includes a first obtaining unit and a first determining unit;

the first obtaining unit is used for obtaining first vehicle passing record data of the vehicle to be analyzed passing through a gate within first preset time;

the first determining unit is used for determining the vehicle track of the vehicle to be analyzed according to a preset track division rule and the first vehicle passing record data.

In a specific implementation manner, the classification module 420 includes an input unit, a second obtaining unit, and a dividing unit;

the input unit is used for inputting the vehicle track of the vehicle to be analyzed into the theme model;

the second obtaining unit is used for obtaining a theme corresponding to each vehicle track output by the theme model;

the dividing unit is used for dividing the vehicle tracks with the same theme into one type.

In a specific implementation manner, the topic model is a latent dirichlet allocation LDA model; the LDA model comprises a sample bayonet identification, a theme and a sample vehicle track three-layer structure of the sample bayonet;

as shown in fig. 5, the vehicle behavior analysis apparatus provided in the embodiment of the present invention may further include a topic model building module for building a topic model, where the topic model building module includes a third obtaining unit 510, a second determining unit 520, a third determining unit 530, and a fourth obtaining unit 540;

the third obtaining unit 510 is configured to obtain sample vehicle-passing record data of each sample vehicle of each sample gate within a second predetermined time, where the sample vehicle-passing record data includes sample identification information of the sample vehicle;

the second determining unit 520 is configured to determine sample vehicle-passing record data corresponding to the sample identification information of each sample vehicle from the sample vehicle-passing record data;

the third determining unit 530 is configured to determine a sample vehicle trajectory of a corresponding sample vehicle according to a preset trajectory division rule and sample vehicle-passing record data corresponding to the same sample identification information, where the sample vehicle trajectory includes a sample gate identification of a sample gate through which the vehicle passes;

the fourth obtaining unit 540 is configured to perform LDA learning, obtain a corresponding relationship between each sample gate identifier and a theme, and obtain a conversion relationship between the theme of the sample gate identifier and the theme of the sample vehicle trajectory;

the second obtaining unit is specifically configured to, for the bayonet identifier of each input vehicle track, determine a theme corresponding to the bayonet identifier according to a correspondence between each sample bayonet identifier and the theme, determine a theme of each vehicle track according to a conversion relationship between the theme of the sample bayonet identifier and the theme of the sample vehicle track, and output the determined theme of each vehicle track; a theme of each of the outputted vehicle trajectories is obtained.

In a specific implementation manner, the first vehicle passing record data includes a first vehicle passing time when the vehicle to be analyzed passes through a corresponding gate;

as shown in fig. 6, the first determination unit includes a first ordering sub-module 610, a first determination sub-module 620, a second determination sub-module 630, and a third determination sub-module 640;

the first sequencing submodule 610 is configured to sequence the first vehicle passing record data according to the corresponding first vehicle passing time sequence;

the first determining submodule 620 is configured to determine a first time difference between every two adjacent first vehicle passing record data according to the first vehicle passing time corresponding to each first vehicle passing record data;

the second determining submodule 630 is configured to determine, when the first time difference exceeds a preset first track division time threshold, two corresponding adjacent first vehicle passing record data as a first track division limit, where a bayonet identifier of a bayonet corresponding to a first vehicle passing record data before a first vehicle passing time is a previous vehicle track end identifier, and a bayonet identifier of a bayonet corresponding to a first vehicle passing record data after the first vehicle passing time is a next vehicle track start identifier;

the third determining submodule 640 is configured to determine the vehicle trajectory of the vehicle to be analyzed according to the first trajectory partition boundary and the first vehicle passing record data.

In a specific implementation manner, the sample vehicle-passing record data further includes a second vehicle-passing time when the corresponding sample vehicle passes through the corresponding gate;

the third determining unit comprises a second sequencing submodule, a fourth determining submodule, a fifth determining submodule and a sixth determining submodule;

the second sequencing submodule is used for sequencing the sample vehicle-passing record data according to the corresponding second vehicle-passing time sequence;

the fourth determining submodule is used for determining a second time difference of every two adjacent sample vehicle-passing record data according to a second vehicle-passing time corresponding to each sample vehicle-passing record data;

the fifth determining submodule is used for taking the corresponding two adjacent sample vehicle-passing record data as a second track dividing boundary when the second time difference exceeds a preset second track dividing time threshold, wherein a bayonet mark of a bayonet corresponding to the vehicle-passing record data before the second vehicle-passing time is a previous vehicle track end mark, and a bayonet mark of a bayonet corresponding to the vehicle-passing record data after the second vehicle-passing time is a next vehicle track start mark;

and the sixth determining submodule is used for determining the vehicle track of the corresponding sample vehicle according to the second track dividing boundary and the sample vehicle passing record data.

In a specific implementation, the bayonet information includes: longitude and latitude information of the bayonet corresponding to each bayonet identification, building information in a preset range of the bayonet corresponding to each bayonet identification or environment information of the bayonet corresponding to each bayonet identification;

the determining module 430 is specifically configured to determine a region theme corresponding to each type of vehicle track according to longitude and latitude information corresponding to the gate identifier included in each vehicle track; determining a region theme corresponding to the vehicle to be analyzed within the first preset time according to the region theme corresponding to each determined vehicle track; or determining a building theme corresponding to each type of vehicle track according to building information in a bayonet preset range corresponding to a bayonet identification contained in each vehicle track; determining a building theme corresponding to the vehicle to be analyzed within the first preset time according to the building theme corresponding to each determined vehicle track; or determining an environment theme corresponding to each type of vehicle track according to environment information of a bayonet corresponding to the bayonet identification contained in each vehicle track; and determining an environment theme corresponding to the vehicle to be analyzed within the first preset time according to the environment theme corresponding to each determined vehicle track.

For the system/apparatus embodiments, since they are substantially similar to the method embodiments, the description is relatively simple, and reference may be made to some descriptions of the method embodiments for relevant points.

It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.

Those skilled in the art will appreciate that all or part of the steps in the above method embodiments may be implemented by a program to instruct relevant hardware to perform the steps, and the program may be stored in a computer-readable storage medium, which is referred to herein as a storage medium, such as: ROM/RAM, magnetic disk, optical disk, etc.

The above description is only for the preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention shall fall within the protection scope of the present invention.

Claims (10)

1. A vehicle behavior analysis method, characterized in that the method comprises:

obtaining a vehicle track of a vehicle to be analyzed in first preset time, wherein the vehicle track comprises a bayonet identification of a bayonet through which the vehicle to be analyzed passes; each vehicle track at least comprises bayonet identifications of two bayonets;

classifying the vehicle track of the vehicle to be analyzed according to a preset theme model;

for each type of vehicle track, determining behavior information of the vehicle to be analyzed in the first preset time according to the gate information corresponding to the gate identification contained in each vehicle track;

wherein the mount information includes: longitude and latitude information of the bayonet corresponding to each bayonet identification, building information in a preset range of the bayonet corresponding to each bayonet identification or environment information of the bayonet corresponding to each bayonet identification;

for each type of vehicle track, determining behavior information of the vehicle to be analyzed in the first preset time according to the gate information corresponding to the gate identifier contained in each vehicle track, including:

determining a region theme corresponding to each type of vehicle track according to longitude and latitude information corresponding to the gate identification contained in each vehicle track; determining a region theme corresponding to the vehicle to be analyzed within the first preset time according to the region theme corresponding to each determined vehicle track, and further analyzing behavior information corresponding to the vehicle to be analyzed; wherein the behavior information includes: the moving region range of the vehicle to be analyzed, or the probability of the vehicle to be analyzed appearing in the designated moving region range;

or the like, or, alternatively,

determining a building theme corresponding to each type of vehicle track according to building information in a bayonet preset range corresponding to a bayonet mark contained in each vehicle track; determining a building theme corresponding to the vehicle to be analyzed within the first preset time according to the building theme corresponding to each determined vehicle track, and further analyzing behavior information corresponding to the vehicle to be analyzed; wherein the behavior information includes: the interest and hobbies of the corresponding personnel of the vehicle to be analyzed, or the range of the moving region of the vehicle to be analyzed;

or the like, or, alternatively,

determining an environment theme corresponding to each type of vehicle track according to environment information of a bayonet corresponding to a bayonet identification contained in each vehicle track; determining an environment theme corresponding to the vehicle to be analyzed within the first preset time according to the determined environment theme corresponding to each type of vehicle track, and further analyzing behavior information corresponding to the vehicle to be analyzed; wherein the behavior information includes: the moving range of the vehicle to be analyzed, or the probability of the vehicle to be analyzed appearing in a specified type of place;

the obtaining of the vehicle track of the vehicle to be analyzed within the first preset time comprises:

obtaining first vehicle passing record data of the vehicle to be analyzed passing through a gate in first preset time;

and determining the vehicle track of the vehicle to be analyzed according to a preset track division rule and the first vehicle passing record data.

2. The method according to claim 1, wherein the classifying the vehicle trajectory of the vehicle to be analyzed according to the preset topic model comprises:

inputting the vehicle track of the vehicle to be analyzed into the theme model;

obtaining a theme corresponding to each vehicle track output by the theme model;

the vehicle trajectories with the same theme are classified into one category.

3. The method of claim 2, wherein the topic model is a latent dirichlet distribution LDA model; the LDA model comprises a sample bayonet identification, a theme and a sample vehicle track three-layer structure of the sample bayonet;

the method further includes a process of building a topic model, the process including:

obtaining sample vehicle-passing record data of each sample vehicle of each sample gate in a second preset time, wherein the sample vehicle-passing record data comprises sample identification information of the sample vehicle;

determining sample vehicle-passing record data corresponding to the sample identification information of each sample vehicle from the sample vehicle-passing record data;

determining a sample vehicle track of a corresponding sample vehicle according to a preset track division rule and sample vehicle-passing record data corresponding to the same sample identification information, wherein the sample vehicle track comprises a sample gate identification of a sample gate through which the vehicle passes;

performing LDA learning to obtain the corresponding relation between each sample gate mark and the theme and the conversion relation between the theme of the sample gate mark and the theme of the sample vehicle track;

the obtaining of the theme corresponding to each vehicle track output by the theme model includes:

the theme model determines a theme corresponding to the bayonet identification according to the corresponding relation between each sample bayonet identification and the theme aiming at the bayonet identification of each input vehicle track, determines the theme of each vehicle track according to the conversion relation between the theme of the sample bayonet identification and the theme of the sample vehicle track, and outputs the determined theme of each vehicle track;

a theme of each of the outputted vehicle trajectories is obtained.

4. The method according to claim 1, wherein the first passing record data comprises a first passing time when the vehicle to be analyzed passes through a corresponding gate;

the determining the vehicle track of the vehicle to be analyzed according to the preset track division rule and the first vehicle passing record data comprises the following steps:

sequencing the first vehicle passing record data according to the corresponding first vehicle passing time sequence;

determining a first time difference of every two adjacent first vehicle passing record data according to the first vehicle passing time corresponding to each first vehicle passing record data;

when the first time difference exceeds a preset first track division time threshold value, determining two corresponding adjacent first vehicle passing record data as a first track division limit, wherein a bayonet mark of a bayonet corresponding to the first vehicle passing record data before the first vehicle passing time is a previous vehicle track end mark, and a bayonet mark of a bayonet corresponding to the first vehicle passing record data after the first vehicle passing time is a next vehicle track start mark;

and determining the vehicle track of the vehicle to be analyzed according to the first track dividing boundary and the first vehicle passing record data.

5. The method of claim 3, wherein the sample vehicle-passing log data further includes a second time-of-pass for the corresponding sample vehicle to pass through the corresponding gate;

the determining the sample vehicle track of the corresponding sample vehicle according to the preset track division rule and the sample vehicle-passing record data corresponding to the same sample identification information includes:

sequencing the sample vehicle-passing record data according to the corresponding second vehicle-passing time sequence;

determining a second time difference of every two adjacent sample vehicle-passing record data according to a second vehicle-passing time corresponding to each sample vehicle-passing record data;

when the second time difference exceeds a preset second track division time threshold value, taking the corresponding two adjacent sample vehicle-passing record data as a second track division boundary, wherein a bayonet mark of a bayonet corresponding to the vehicle-passing record data before the second vehicle-passing time is a previous vehicle track end mark, and a bayonet mark of a bayonet corresponding to the vehicle-passing record data after the second vehicle-passing time is a next vehicle track start mark;

and determining the vehicle track of the corresponding sample vehicle according to the second track dividing boundary and the sample vehicle passing record data.

6. A vehicle behavior analysis apparatus characterized by comprising:

the analysis method comprises the steps of obtaining a vehicle track of a vehicle to be analyzed in first preset time, wherein the vehicle track comprises a bayonet identification of a bayonet through which the vehicle to be analyzed passes; each vehicle track at least comprises bayonet identifications of two bayonets;

the classification module is used for classifying the vehicle track of the vehicle to be analyzed according to a preset theme model;

the determining module is used for determining behavior information of the vehicle to be analyzed in the first preset time according to the vehicle track of each type and the vehicle track information corresponding to the vehicle gate identification contained in each vehicle track;

wherein the mount information includes: longitude and latitude information of the bayonet corresponding to each bayonet identification, building information in a preset range of the bayonet corresponding to each bayonet identification or environment information of the bayonet corresponding to each bayonet identification; the determining module is specifically configured to:

determining a region theme corresponding to each type of vehicle track according to longitude and latitude information corresponding to the gate identification contained in each vehicle track; determining a region theme corresponding to the vehicle to be analyzed within the first preset time according to the region theme corresponding to each determined vehicle track, and further analyzing behavior information corresponding to the vehicle to be analyzed; wherein the behavior information includes: the moving region range of the vehicle to be analyzed, or the probability of the vehicle to be analyzed appearing in the designated moving region range;

or the like, or, alternatively,

determining a building theme corresponding to each type of vehicle track according to building information in a bayonet preset range corresponding to a bayonet mark contained in each vehicle track; determining a building theme corresponding to the vehicle to be analyzed within the first preset time according to the building theme corresponding to each determined vehicle track, and further analyzing behavior information corresponding to the vehicle to be analyzed; wherein the behavior information includes: the interest and hobbies of the corresponding personnel of the vehicle to be analyzed, or the range of the moving region of the vehicle to be analyzed;

or the like, or, alternatively,

determining an environment theme corresponding to each type of vehicle track according to environment information of a bayonet corresponding to a bayonet identification contained in each vehicle track; determining an environment theme corresponding to the vehicle to be analyzed within the first preset time according to the determined environment theme corresponding to each type of vehicle track, and further analyzing behavior information corresponding to the vehicle to be analyzed; wherein the behavior information includes: the moving range of the vehicle to be analyzed, or the probability of the vehicle to be analyzed appearing in a specified type of place;

wherein the obtaining module comprises a first obtaining unit and a first determining unit;

the first obtaining unit is used for obtaining first vehicle passing record data of the vehicle to be analyzed passing a gate within first preset time, wherein the first vehicle passing record data comprises identification information of the corresponding vehicle;

the first determining unit is used for determining the vehicle track of the vehicle to be analyzed according to a preset track division rule and the first vehicle passing record data.

7. The apparatus of claim 6, wherein the classification module comprises an input unit, a second obtaining unit, and a dividing unit;

the input unit is used for inputting the vehicle track of the vehicle to be analyzed into the theme model;

the second obtaining unit is used for obtaining a theme corresponding to each vehicle track output by the theme model;

the dividing unit is used for dividing the vehicle tracks with the same theme into one type.

8. The apparatus of claim 7, wherein the topic model is a latent dirichlet distribution LDA model; the LDA model comprises a sample bayonet identification, a theme and a sample vehicle track three-layer structure of the sample bayonet;

the device also comprises a theme model establishing module used for establishing a theme model, wherein the theme model establishing module comprises a third obtaining unit, a second determining unit, a third determining unit and a fourth obtaining unit;

the third obtaining unit is used for obtaining sample vehicle-passing record data of each sample vehicle of each sample gate in a second preset time, wherein the sample vehicle-passing record data comprises sample identification information of the sample vehicle;

the second determining unit is used for determining sample vehicle-passing record data corresponding to the sample identification information of each sample vehicle from the sample vehicle-passing record data;

the third determining unit is used for determining a sample vehicle track of the corresponding sample vehicle according to a preset track division rule and sample vehicle-passing record data corresponding to the same sample identification information, wherein the sample vehicle track comprises a sample gate identification of a sample gate through which the vehicle passes;

the fourth obtaining unit is used for performing LDA learning to obtain a corresponding relation between each sample gate identifier and the theme and a conversion relation between the theme of the sample gate identifier and the theme of the sample vehicle track;

the second obtaining unit is specifically configured to, for the bayonet identifier of each input vehicle track, determine a theme corresponding to the bayonet identifier according to a correspondence between each sample bayonet identifier and the theme, determine a theme of each vehicle track according to a conversion relationship between the theme of the sample bayonet identifier and the theme of the sample vehicle track, and output the determined theme of each vehicle track;

a theme of each of the outputted vehicle trajectories is obtained.

9. The apparatus of claim 6, wherein the first passing record data comprises a first passing time when the vehicle to be analyzed passes through the corresponding gate;

the first determining unit comprises a first sequencing submodule, a first determining submodule, a second determining submodule and a third determining submodule;

the first sequencing submodule is used for sequencing the first vehicle passing record data according to the corresponding first vehicle passing time sequence;

the first determining submodule is used for determining a first time difference between every two adjacent first vehicle passing record data according to the first vehicle passing time corresponding to each first vehicle passing record data;

the second determining submodule is used for determining two corresponding adjacent first vehicle passing record data as a first track dividing boundary when the first time difference exceeds a preset first track dividing time threshold, wherein a bayonet mark of a bayonet corresponding to the first vehicle passing record data before the first vehicle passing time is a previous vehicle track end mark, and a bayonet mark of a bayonet corresponding to the first vehicle passing record data after the first vehicle passing time is a next vehicle track start mark;

the third determining submodule is used for determining the vehicle track of the vehicle to be analyzed according to the first track dividing boundary and the first vehicle passing record data.

10. The apparatus of claim 8, wherein the sample vehicle-passing log data further comprises a second time-of-pass for the corresponding sample vehicle to pass through the corresponding gate;

the third determining unit comprises a second sequencing submodule, a fourth determining submodule, a fifth determining submodule and a sixth determining submodule;

the second sequencing submodule is used for sequencing the sample vehicle-passing record data according to the corresponding second vehicle-passing time sequence;

the fourth determining submodule is used for determining a second time difference of every two adjacent sample vehicle-passing record data according to a second vehicle-passing time corresponding to each sample vehicle-passing record data;

the fifth determining submodule is used for taking the corresponding two adjacent sample vehicle-passing record data as a second track dividing boundary when the second time difference exceeds a preset second track dividing time threshold, wherein a bayonet mark of a bayonet corresponding to the vehicle-passing record data before the second vehicle-passing time is a previous vehicle track end mark, and a bayonet mark of a bayonet corresponding to the vehicle-passing record data after the second vehicle-passing time is a next vehicle track start mark;

and the sixth determining submodule is used for determining the vehicle track of the corresponding sample vehicle according to the second track dividing boundary and the sample vehicle passing record data.

Images