CN107862862A

CN107862862A - A kind of vehicle behavior analysis method and device

Info

Publication number: CN107862862A
Application number: CN201610840135.8A
Authority: CN
Inventors: 何春雷; 管国辰; 林武康; 徐宇垚
Original assignee: Hangzhou Hikvision Digital Technology Co Ltd
Current assignee: Hangzhou Hikvision Digital Technology Co Ltd
Priority date: 2016-09-22
Filing date: 2016-09-22
Publication date: 2018-03-30
Anticipated expiration: 2036-09-22
Also published as: CN107862862B

Abstract

The embodiment of the invention discloses a kind of vehicle behavior analysis method and device, this method to include：Track of vehicle of the vehicle to be analyzed within first scheduled time is obtained, wherein, the track of vehicle includes the bayonet socket mark for the bayonet socket that the vehicle to be analyzed is passed through；According to default topic model, the track of vehicle of the vehicle to be analyzed is classified；To every a kind of track of vehicle, according to bayonet socket information corresponding to the bayonet socket mark that every track of vehicle is included, behavioural information of the vehicle to be analyzed within first scheduled time is determined.It can be seen that according to default topic model, the inner link of track of vehicle of the vehicle to be analyzed within first scheduled time is determined, further, according to the inner link of track of vehicle, determines behavioural information of the vehicle to be analyzed within first scheduled time.Follow-up is tracked according to behavior information to the vehicle to be analyzed, can improve the accuracy rate of the tracking to the vehicle to be analyzed.

Description

A kind of vehicle behavior analysis method and device

Technical field

The present invention relates to Data Mining, more particularly to a kind of vehicle behavior analysis method and device.

Background technology

Current era, vehicle increasingly become the walking-replacing tool of people.It is substantial amounts of collected by monitoring device on road Information of vehicles is (such as：Cross car record etc.), typically it is intended merely to search violation vehicle, does not pay close attention to big collected by each monitoring device Implicit information in the information of vehicles of amount.Such as：Substantial amounts of information of vehicles according to collected by monitoring device, determine the vehicle of vehicle Track, etc..

Because current era vehicle increasingly becomes the walking-replacing tool of people, the track of vehicle of vehicle is further analyzed, can Further to determine the behavioural information of vehicle, the further some relevant informations for determining driver, such as the interest of personnel The information such as hobby, behavioural habits.

So, how for the substantial amounts of information of vehicles collected by the monitoring device on road, behavior point is carried out to vehicle Analysis turns into urgent problem to be solved.

The content of the invention

The embodiment of the invention discloses a kind of vehicle behavior analysis method and device, to realize the vehicle rail according to vehicle Mark, obtain the behavioural information of corresponding vehicle.Concrete scheme is as follows：

On the one hand, the embodiments of the invention provide a kind of vehicle behavior analysis method, methods described to include：

Track of vehicle of the vehicle to be analyzed within first scheduled time is obtained, wherein, the track of vehicle includes described treat The bayonet socket mark for the bayonet socket that analysis vehicle is passed through；

According to default topic model, the track of vehicle of the vehicle to be analyzed is classified；

To every a kind of track of vehicle, according to bayonet socket information corresponding to the bayonet socket mark that every track of vehicle is included, it is determined that Behavioural information of the vehicle to be analyzed within first scheduled time.

Optionally, the track of vehicle for obtaining vehicle to be analyzed within first scheduled time, including：

Obtain the vehicle to be analyzed and cross car record data by the first of bayonet socket within first scheduled time；

Car record data is crossed according to default track division rule and described first, determines the car of the vehicle to be analyzed Track.

Optionally, it is described according to default topic model, the track of vehicle of the vehicle to be analyzed is classified, including：

The track of vehicle of the vehicle to be analyzed is inputted into the topic model；

Obtain theme corresponding to every track of vehicle of the topic model output；

Theme identical track of vehicle is divided into one kind.

Optionally, the topic model is that implicit Di Li Crays are distributed LDA models；The LDA models include sample bayonet socket Sample bayonet socket mark, theme and sample track of vehicle three-decker；

Methods described also includes the process for establishing topic model, and the process includes：

The sample for obtaining each sample vehicle of each sample bayonet socket within second scheduled time crosses car record data, its In, the sample, which crosses car record data, includes the sample identification information of sample vehicle；

Crossed from the sample in car record data, it is determined that sample corresponding to the sample identification information of each sample vehicle crosses car Record data；

Car record data is crossed according to the sample of default track division rule and corresponding same sample identification information, really The sample track of vehicle of fixed corresponding sample vehicle, the sample track of vehicle include the sample for the sample bayonet socket that vehicle is passed through This bayonet socket identifies；

LDA study is carried out, obtains each sample bayonet socket mark and the corresponding relation of theme, and the master of sample bayonet socket mark Transformational relation between the theme of topic and sample track of vehicle；

Theme corresponding to every track of vehicle for obtaining the topic model output, including：

The topic model identifies and main for the bayonet socket mark of every track of vehicle of input according to each sample bayonet socket The corresponding relation of topic determines theme corresponding to bayonet socket mark, and according to the master of the theme that sample bayonet socket identifies and sample track of vehicle Transformational relation between topic determines the theme of every track of vehicle, the theme of every track of vehicle determined by output；

Obtain the theme of the every track of vehicle exported.

Optionally, described first when crossing car record data and including the vehicle to be analyzed and cross car by the first of corresponding bayonet socket Between；

It is described according to default track division rule and it is described first cross car record data, determine the vehicle to be analyzed Track of vehicle, including：

Car time order and function order is crossed according to corresponding first, car record data is crossed in sequence described first；

The car time is spent according to first corresponding to each first excessively car record data, determines the first adjacent mistake car of each two The very first time of record data is poor；

When very first time difference exceedes default first track division time threshold, corresponding two are adjacent First car record data excessively is defined as the first track division limits, wherein, first crosses the car record number excessively of car time preceding first Previous track of vehicle terminal mark is identified as according to the bayonet socket of corresponding bayonet socket, the first car time posterior first excessively crossed car record The bayonet socket of bayonet socket corresponding to data is identified as the latter track of vehicle and plays point identification；

Car record data is crossed according to first track division limits and described first, determines the vehicle to be analyzed Track of vehicle.

Optionally, the sample crosses car record data and also crosses car by the second of corresponding bayonet socket including corresponding sample vehicle Time；

The sample according to default track division rule and corresponding same sample identification information crosses car record number According to, it is determined that the sample track of vehicle of corresponding sample vehicle, including：

Car time order and function order is crossed according to corresponding second, the sample that sorts crosses car record data；

The second mistake car time corresponding to car record data is spent according to each sample, determines that the adjacent sample of each two crosses car Second time difference of record data；

When second time difference exceeding default second track division time threshold, corresponding two are adjacent Sample crosses car record data as the second track division limits, wherein, second, which crosses car time preceding car record data of crossing, corresponds to The bayonet socket of bayonet socket be identified as previous track of vehicle terminal mark, second, which spends the car time, posterior crosses corresponding to car record data The bayonet socket of bayonet socket is identified as the latter track of vehicle and plays point identification；

Car record data is crossed according to second track division limits and the sample, it is determined that corresponding sample vehicle Track of vehicle.

Optionally, the bayonet socket information includes：The latitude and longitude information of bayonet socket, each bayonet socket mark pair corresponding to each bayonet socket mark Environmental information corresponding to the architecture information in bayonet socket preset range answered or each bayonet socket mark residing for bayonet socket；

It is described to every a kind of track of vehicle, according to bayonet socket information corresponding to the bayonet socket mark that every track of vehicle is included, Behavioural information of the vehicle to be analyzed within first scheduled time is determined, including：

Determined according to latitude and longitude information corresponding to the bayonet socket mark that every track of vehicle is included per a kind of track of vehicle pair The region theme answered；

According to region theme corresponding to identified every a kind of track of vehicle, determine the vehicle to be analyzed described first Corresponding region theme in the scheduled time；Or,

Determined according to the architecture information corresponding to the bayonet socket mark that every track of vehicle is included in bayonet socket preset range every Building theme corresponding to a kind of track of vehicle；

Theme is built according to corresponding to identified every a kind of track of vehicle, determines the vehicle to be analyzed described first Corresponding building theme in the scheduled time；Or,

According to the environmental information residing for bayonet socket corresponding to the bayonet socket mark that every track of vehicle is included, it is determined that per a kind of car Environment theme corresponding to track；

According to environment theme corresponding to identified every a kind of track of vehicle, determine the vehicle to be analyzed described first Corresponding environment theme in the scheduled time.

On the other hand, the embodiments of the invention provide a kind of vehicle behavioural analysis device, described device to include：

Module is obtained, for obtaining track of vehicle of the vehicle to be analyzed within first scheduled time, wherein, the vehicle rail Mark includes the bayonet socket mark for the bayonet socket that the vehicle to be analyzed is passed through；

Sort module, for according to default topic model, classifying to the track of vehicle of the vehicle to be analyzed；

Determining module, for every a kind of track of vehicle, according to corresponding to the bayonet socket mark that every track of vehicle is included Bayonet socket information, determine behavioural information of the vehicle to be analyzed within first scheduled time.

Optionally, the acquisition module includes first obtains unit and the first determining unit；

The first obtains unit, for obtain the vehicle to be analyzed within first scheduled time by the of bayonet socket One crosses car record data, wherein, described first, which crosses car record data, includes the identification information of corresponding vehicle；

First determining unit, for crossing car record data according to default track division rule and described first, Determine the track of vehicle of the vehicle to be analyzed.

Optionally, the sort module includes input block, the second obtaining unit and division unit；

The input block, for the track of vehicle of the vehicle to be analyzed to be inputted into the topic model；

Second obtaining unit, for obtaining theme corresponding to every track of vehicle of the topic model output；

The division unit, for theme identical track of vehicle to be divided into one kind.

The topic model that described device also includes being used to establish topic model establishes module, and the topic model establishes module Including the 3rd obtaining unit, the second determining unit, the 3rd determining unit and the 4th obtaining unit；

3rd obtaining unit, for obtaining each sample vehicle of each sample bayonet socket within second scheduled time Sample crosses car record data, wherein, the sample, which crosses car record data, includes the sample identification information of sample vehicle；

Second determining unit, for being crossed from the sample in car record data, it is determined that the sample of each sample vehicle Sample corresponding to identification information crosses car record data；

3rd determining unit, for according to default track division rule and corresponding same sample identification information Sample cross car record data, it is determined that the sample track of vehicle of corresponding sample vehicle, the sample track of vehicle includes car The sample bayonet socket mark of the sample bayonet socket passed through；

4th obtaining unit, for carrying out LDA study, obtain each sample bayonet socket mark pass corresponding with theme Transformational relation between system, and the theme of sample bayonet socket mark and the theme of sample track of vehicle；

Second obtaining unit, the bayonet socket mark specifically for the topic model for every track of vehicle of input Know, identified according to each sample bayonet socket and determine the corresponding theme of bayonet socket mark with the corresponding relation of theme, and according to sample bayonet socket Transformational relation between the theme of mark and the theme of sample track of vehicle determines the theme of every track of vehicle, determined by output The theme of every track of vehicle；

Obtain the theme of the every track of vehicle exported.

First determining unit includes the first sorting sub-module, the first determination sub-module, the second determination sub-module and the Three determination sub-modules；

First sorting sub-module, for crossing car time order and function order, sequence described first according to corresponding first Cross car record data；

First determination sub-module, for spending the car time according to first corresponding to each first excessively car record data, Determine that the very first time of the first adjacent car record data excessively of each two is poor；

Second determination sub-module, for exceeding default first track division time threshold when very first time difference When, by corresponding two it is adjacent first cross car record data be defined as the first track division limits, wherein, first cross car when Between the preceding first bayonet socket for crossing bayonet socket corresponding to car record data be identified as previous track of vehicle terminal mark, first crosses car The bayonet socket that time posterior first crosses bayonet socket corresponding to car record data is identified as the latter track of vehicle and plays point identification；

3rd determination sub-module, for crossing car record number according to first track division limits and described first According to determining the track of vehicle of the vehicle to be analyzed.

3rd determining unit includes the second sorting sub-module, the 4th determination sub-module, the 5th determination sub-module and the Six determination sub-modules；

Second sorting sub-module, for crossing car time order and function order according to corresponding second, sort the sample Cross car record data；

4th determination sub-module, for spending the second mistake car time corresponding to car record data according to each sample, Determine that it's the second time difference of car record data pasts the adjacent sample of each two；

5th determination sub-module, for exceeding default second track division time threshold when second time difference When, two corresponding adjacent samples are crossed into car record data as the second track division limits, wherein, second spends the car time The preceding bayonet socket of bayonet socket excessively corresponding to car record data is identified as previous track of vehicle terminal mark, and the second car time excessively existed The bayonet socket for crossing bayonet socket corresponding to car record data afterwards is identified as the latter track of vehicle and plays point identification；

6th determination sub-module, for crossing car record number according to second track division limits and the sample According to it is determined that the track of vehicle of corresponding sample vehicle.

The determining module, specifically for included according to every track of vehicle bayonet socket mark corresponding to latitude and longitude information It is determined that region theme corresponding to per a kind of track of vehicle；

In this programme, track of vehicle of the vehicle to be analyzed within first scheduled time is obtained, wherein, the track of vehicle bag Include the bayonet socket mark for the bayonet socket that the vehicle to be analyzed is passed through；According to default topic model, to the car of the vehicle to be analyzed Track is classified；To every a kind of track of vehicle, according to bayonet socket letter corresponding to the bayonet socket mark that every track of vehicle is included Breath, determines behavioural information of the vehicle to be analyzed within first scheduled time.It can be seen that according to default topic model, it is determined that treating point The inner link of track of vehicle of the vehicle within first scheduled time is analysed, further, according to the inner link of track of vehicle, really Fixed behavioural information of the vehicle to be analyzed within first scheduled time.Follow-up can be according to behavior information to the vehicle to be analyzed It is tracked, the accuracy rate of the tracking to the vehicle to be analyzed can be improved.Certainly, any product or method of the present invention is implemented It must be not necessarily required to reach all the above advantage simultaneously.

Brief description of the drawings

In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing There is the required accompanying drawing used in technology description to be briefly described, it should be apparent that, drawings in the following description are only this Some embodiments of invention, for those of ordinary skill in the art, on the premise of not paying creative work, can be with Other accompanying drawings are obtained according to these accompanying drawings.

A kind of schematic flow sheet for vehicle behavior analysis method that Fig. 1 is provided by the embodiment of the present invention；

Fig. 2A is the flow example figure that track of vehicle divides in the embodiment of the present invention；

Fig. 2 B are the exemplary plot that track of vehicle divides in the embodiment of the present invention；

Fig. 3 is a kind of schematic flow sheet that topic model is established in the embodiment of the present invention；

A kind of structural representation for vehicle behavioural analysis device that Fig. 4 is provided by the embodiment of the present invention；

Fig. 5 is the structural representation that topic model establishes module in the embodiment of the present invention；

Fig. 6 is the topology example figure of track of vehicle division module in the embodiment of the present invention.

Embodiment

Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete Site preparation describes, it is clear that described embodiment is only part of the embodiment of the present invention, rather than whole embodiments.It is based on Embodiment in the present invention, those of ordinary skill in the art are obtained every other under the premise of creative work is not made Embodiment, belong to the scope of protection of the invention.

The embodiments of the invention provide a kind of vehicle behavior analysis method and device, to realize the vehicle rail according to vehicle Mark, obtain the behavioural information of corresponding vehicle.

A kind of vehicle behavior analysis method provided first below the embodiment of the present invention is introduced.

As shown in figure 1, the embodiments of the invention provide a kind of vehicle behavior analysis method, may include steps of：

S101：Track of vehicle of the vehicle to be analyzed within first scheduled time is obtained, wherein, the track of vehicle includes institute State the bayonet socket mark for the bayonet socket that vehicle to be analyzed is passed through；

It is understood that the bayonet socket mark of two bayonet sockets can be comprised at least in every track of vehicle of the vehicle to be analyzed Know, and first scheduled time can be half a year or 1 year etc., and the embodiment of the present invention is not entered to first scheduled time Row limits.Wherein it is determined that the mode of track of vehicle is varied, vehicle rail is such as determined according to the location information of corresponding vehicle Mark.

In a kind of specific implementation, the time sequencing of bayonet socket can be passed through according to the vehicle to be analyzed and is passed through The bayonet socket mark of bayonet socket is crossed, determines the track of vehicle of the vehicle to be analyzed, it is described to obtain vehicle to be analyzed in first scheduled time Interior track of vehicle (S101), can include：

Obtain vehicle to be analyzed and cross car record data by the first of bayonet socket within first scheduled time；

It should be noted that corresponding video monitoring equipment or scanning device be present at each bayonet socket, pass through the video Monitoring device or scanning device can record the car record data excessively of the vehicle of the corresponding bayonet socket of each process, wherein, this crosses car note The identification information that can distinguish different vehicle is included in record data.Wherein, obtaining the first car record data excessively can use now There is technology, will not be described here.

In a kind of specific implementation, described first, which crosses car record data, includes the vehicle to be analyzed by corresponding card It's the car time pasts the first of mouth；

As shown in Figure 2 A, it is described according to default track division rule and it is described first cross car record data, determine institute The track of vehicle of vehicle to be analyzed is stated, can be included：

S201：Car time order and function order is crossed according to corresponding first, car record data is crossed in sequence described first；

S202：Crossed according to each first and first spend the car time corresponding to car record data, determine each two it is adjacent the The very first time of one car record data excessively is poor；

S203：When very first time difference exceedes default first track division time threshold, by corresponding two The car record data excessively of adjacent first is defined as the first track division limits, wherein, first crosses the car excessively of car time preceding first The bayonet socket of bayonet socket corresponding to record data is identified as previous track of vehicle terminal mark, and first crosses car time posterior first mistake The bayonet socket of bayonet socket corresponding to car record data is identified as the latter track of vehicle and plays point identification；

S204：Car record data is crossed according to first track division limits and described first, is determined described to be analyzed The track of vehicle of vehicle.

Wherein, the vehicle that video monitoring equipment corresponding to each bayonet socket is recorded cross car record data in, except including There is the identification information of corresponding vehicle, the car time excessively of correspondingly vehicle can also be included, specifically, the first mistake car record data Include vehicle to be analyzed and spend the car time by corresponding to the first of bayonet socket.

Pass through the first mistake car time, it may be determined that the vehicle to be analyzed passes through the sequencing of each bayonet socket, Jin Erke To determine the track of vehicle of vehicle to be analyzed.It is understood that car is crossed by the first of each bayonet socket according to vehicle to be analyzed Record data, the division of track of vehicle is carried out, can be determined according to vehicle to be analyzed in some local residence time.Typically In the case of, if vehicle is constantly in transport condition, it is generally a few minutes to tens by the time interval of each two bayonet socket Minute or so.In other words, can be true when vehicle passes through the time interval relatively large (such as more than two hours) of two bayonet sockets The fixed vehicle stops between two bayonet sockets, and vehicle is bigger by the time interval of two bayonet sockets, may indicate that the car Residence time between two bayonet sockets is longer.At this point it is possible to corresponding two bayonet sockets are divided as the first track Boundary, it is whole that the first bayonet socket that the car time preceding first crosses bayonet socket corresponding to car record data excessively is identified as previous track of vehicle Point identification, the first bayonet socket that the car time posterior first crosses bayonet socket corresponding to car record data excessively are identified as the latter track of vehicle Play point identification.

For example, as shown in Figure 2 B, the exemplary plot that the track of vehicle provided by the embodiment of the present invention divides；In Fig. 2 B Shown coordinate is the time scale of one day, and mark needs the first mistake for passing through bayonet socket in being analyzed vehicle A mono- day on the coordinate Car record data, wherein, first, which crosses record in car record data, the bayonet socket mark and the vehicle A to be analyzed of corresponding bayonet socket Identification information, specifically include：19 divide the-bayonet socket mark of bayonet socket 1 during A-6,19 divide (first when being embodied as vehicle A6 to be analyzed Spend the car time) 1 divide the-bayonet socket mark of bayonet socket 2 when passing through bayonet socket 1, A-7,1 divides (first when being embodied as vehicle A7 to be analyzed Spend the car time) 50 divide the-bayonet socket mark of bayonet socket 3 when passing through bayonet socket 2, A-7,50 divide when being embodied as vehicle A7 to be analyzed (the One spends the car time) 50 divide when pass through bayonet socket 3, A-11-bayonet socket of bayonet socket 4 identifies, 50 divide when being embodied as vehicle A11 to be analyzed 40 divide when (first cross car time) pass through bayonet socket 4, A-12-bayonet socket of bayonet socket 5 identifies, when being embodied as vehicle A12 to be analyzed 28 divide when 40 points (first cross car times) pass through bayonet socket 5, A-13-bayonet socket of bayonet socket 6 identifies, it is embodied as vehicle to be analyzed 20 divide when 28 dividing (first spends the car time) pass through bayonet socket 6, A-17 during A13-bayonet socket of bayonet socket 3 identifies, it is embodied as to be analyzed 58 divide when 20 dividing (first spends the car time) pass through bayonet socket 3, A-17 during vehicle A17-bayonet socket of bayonet socket 2 identifies, it is embodied as treating 52 divide when 58 dividing (first spends the car time) pass through bayonet socket 2 and A-18 when analyzing vehicle A17-bayonet socket of bayonet socket 1 identifies, specific table 52 divide (first spends the car time) to pass through bayonet socket 1 when being shown as vehicle A18 to be analyzed；Wherein, default first track division time threshold It is worth for 2 hours.

The very first time that vehicle A to be analyzed is crossed between the temporally adjacent bayonet socket of car by each two corresponding first is calculated respectively Difference.Each very first time difference calculated is contrasted with the first track division time threshold (2 hours).Wherein it is possible to really Surely it is (when 7 by bayonet socket 3 and by the very first time difference between bayonet socket 4 more than the first track division time threshold (2 hours) 50 divide at 50 point -11) 4 hours；By bayonet socket 6 and by the very first time difference between bayonet socket 3 time is divided more than the first track Threshold value (2 hours), for 52 points of (28 dividing when 20 divide -13 when 17) 3 hours；Respectively by above-mentioned two corresponding the first adjacent mistakes Car record data is defined as the first track division limits.

Specifically, crossing car record data according to above-mentioned first, it is respectively 1-2-3 that can be divided into three tracks of vehicle；4- 5-6；3-2-1.

S102：According to default topic model, the track of vehicle of the vehicle to be analyzed is classified；

It is understood that pass through the bayonet socket mark for training substantial amounts of bayonet socket and its corresponding species in the topic model And the track of vehicle identified including bayonet socket and its corresponding species are learnt, the training learnt according to the topic model As a result, the track of vehicle of the vehicle to be analyzed can be classified.Wherein, the topic model can use prior art LDA (LatentDirichletallocation, implying the distribution of Di Li Crays) model.

It is described according to default topic model in a kind of specific implementation, to the track of vehicle of the vehicle to be analyzed Classified, can be included：

Theme identical track of vehicle is divided into one kind.

It should be noted that the topic model can continue being somebody's turn to do for training input according to the training result learnt in advance The track of vehicle of vehicle to be analyzed, theme corresponding to every track of vehicle is trained, and will corresponding theme identical track of vehicle One kind is divided into, wherein, according to existing topic model technology, every track of vehicle can belong in inhomogeneity theme simultaneously, Probability of certain a kind of theme may be similar and different where it.I.e. every track of vehicle may correspond to a kind of or multiclass theme. The topic model is used for the implicit associations based on input information, it is determined that input information belongs to the probability of a certain theme, such as input letter Cease for track of vehicle when, identified comprising bayonet socket in track of vehicle, bayonet socket mark is the implicit associations between track of vehicle.Enter one Step, the probability of a certain theme is belonged to according to input information, the input information is clustered or classified.

S103：To every a kind of track of vehicle, according to bayonet socket information corresponding to the bayonet socket mark that every track of vehicle is included, Determine behavioural information of the vehicle to be analyzed within first scheduled time.

Wherein, for the vehicle to be analyzed, its corresponding track of vehicle within first scheduled time can exist more Bar.According to every a kind of track of vehicle, theme corresponding to the vehicle to be analyzed is determined.Wherein it is possible to understand, the car to be analyzed Track of vehicle corresponding to may belong to inhomogeneous theme, and the probability that corresponding track of vehicle belongs to a certain theme is bigger Or the track of vehicle of a certain theme account for obtain all tracks of vehicle of the vehicle to be analyzed ratio it is bigger, the vehicle to be analyzed point The probability for belonging to a certain theme is bigger.It is made up of due to each track of vehicle the bayonet socket mark of at least two bayonet sockets, root Bayonet socket information corresponding to the bayonet socket mark included according to every track of vehicle, can be that each class theme limits specific implication, Further, can continue to determine behavioural information of the vehicle to be analyzed within first scheduled time.It is understood that the behavior Information can include the probability size that the vehicle to be analyzed appears in certain a kind of place within first scheduled time；Or appear in Some region of probability size.

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

In a kind of specific implementation, the bayonet socket information can include：The longitude and latitude of bayonet socket corresponding to each bayonet socket mark Ring corresponding to architecture information or each bayonet socket mark corresponding to degree information, each bayonet socket mark in bayonet socket preset range residing for bayonet socket Environment information；

It is described to every a kind of track of vehicle, according to bayonet socket information corresponding to the bayonet socket mark that every track of vehicle is included, Behavioural information (S103) of the vehicle to be analyzed within first scheduled time is determined, including：

Prior art, the bayonet socket information according to corresponding to identifying bayonet socket can be used, it may be determined that the corresponding theme of bayonet socket mark Concrete meaning, it is possible to understand that, when bayonet socket information change corresponding to bayonet socket mark, the tool of the corresponding theme of bayonet socket mark Body implication can also change, as corresponding to bayonet socket information is bayonet socket mark during the latitude and longitude information of bayonet socket, what topic model was exported Type of theme corresponds to region theme, and the region theme can include state category or province classification etc., further, analysis The behavioural information of the vehicle to be analyzed, behavior information can be the vehicle to be analyzed often movable territorial scope, or occur In the probability size of a certain movable territorial scope.Building letter as corresponding to bayonet socket information is bayonet socket mark in bayonet socket preset range During breath, type of theme that topic model is exported corresponds to build theme, the building theme can include antique building theme and Modern architecture theme etc., further, analyze hobby, activity venue scope of the vehicle to be analyzed corresponding (personnel) etc.. As corresponding to bayonet socket information is bayonet socket mark during environmental information residing for bayonet socket, the type of theme that topic model is exported can be right The themes such as amusement, food and drink, motion and social activity are should be, further, analyze behavioural information corresponding to the vehicle to be analyzed, behavior letter Breath can include band analysis vehicle and appear in certain a kind of place (such as：Amusement, food and drink, motion and social activity) probability or movable model Enclose.

It is understood that follow-up be able to can improve to be analyzed to this according to the behavioural information of the vehicle to be analyzed The accuracy rate of the tracking of vehicle.Such as：It is that when entertaining, may indicate that the vehicle to be analyzed to determine theme corresponding to vehicle to be analyzed At some bayonet sockets related entertainment theme that can be frequent；The probability meeting of the vehicle to be analyzed is traced into public place of entertainment It is larger.It is other, when it is determined that the vehicle to be analyzed frequently occurs on a certain bayonet socket, very maximum probability it can deduce this and treat point Analyse certain Theme activity that vehicle is carried out.

Using the embodiment of the present invention, track of vehicle of the vehicle to be analyzed within first scheduled time is obtained, wherein, the vehicle Track includes the bayonet socket mark for the bayonet socket that the vehicle to be analyzed is passed through；According to default topic model, to the car to be analyzed Track of vehicle classified；To every a kind of track of vehicle, according to corresponding to the bayonet socket mark that every track of vehicle is included Bayonet socket information, determine behavioural information of the vehicle to be analyzed within first scheduled time.It can be seen that according to default topic model, really The inner link of fixed track of vehicle of the vehicle to be analyzed within first scheduled time, further, according to the inherence of track of vehicle Contact, determines behavioural information of the vehicle to be analyzed within first scheduled time.Follow-up can treat according to behavior information to this Analysis vehicle is tracked, and can improve the accuracy rate of the tracking to the vehicle to be analyzed.

In a kind of specific implementation, the topic model is that implicit Di Li Crays are distributed LDA models；The LDA moulds Type includes sample bayonet socket mark, theme and the sample track of vehicle three-decker of sample bayonet socket；

As shown in figure 3, a kind of vehicle behavior analysis method that the embodiment of the present invention is provided also includes establishing topic model Process, the process can include：

S301：The sample for obtaining each sample vehicle of each sample bayonet socket within second scheduled time crosses car record number According to, wherein, the sample, which crosses car record data, includes the sample identification information of sample vehicle；

S302：Crossed from the sample in car record data, it is determined that sample corresponding to the sample identification information of each sample vehicle This crosses car record data；

S303：Car record number is crossed according to the sample of default track division rule and corresponding same sample identification information According to it is determined that the sample track of vehicle of corresponding sample vehicle, the sample track of vehicle includes the sample bayonet socket that vehicle is passed through Sample bayonet socket mark；

S304：LDA study is carried out, obtains each sample bayonet socket mark and the corresponding relation of theme, and sample bayonet socket mark Transformational relation between the theme of knowledge and the theme of sample track of vehicle；

Obtain the theme of the every track of vehicle exported.

The LDA models include sample bayonet socket mark, theme and the sample track of vehicle three-decker of sample bayonet socket, wherein Corresponding relation can be expressed as sample bayonet socket mark-theme, theme-sample track of vehicle and sample bayonet socket mark-sample Track of vehicle, it is to be understood that every sample track of vehicle is at least two sample bayonet sockets mark composition, passes through sample card Theme corresponding to mouth mark can determine theme corresponding to sample track of vehicle.

The sample for obtaining each sample vehicle within second scheduled time crosses car record data, and according to default rail The sample of mark division rule and each sample vehicle crosses car record data, determines sample vehicle corresponding to each sample vehicle Behind track, LDA study is carried out to sample track of vehicle, that is, starts to train the sample track of vehicle.Wherein it is possible to understand, What the corresponding relation of sample bayonet socket mark-sample track of vehicle was to determine, i.e. bayonet socket mark belongs to a certain sample track of vehicle Probability, also, a certain sample track of vehicle has what the probability that a certain bayonet socket identifies was to determine.Predetermined quantity can be pre-set Individual theme, such as theme 1, theme 2 ... theme N-1, theme N.Same as the prior art, LDA models are sample card at random first Mouth mark-theme and theme-sample track of vehicle carry out assignment, and constantly repeat the process of the assignment.Until sample bayonet socket After mark-theme and theme-numerical value corresponding to sample track of vehicle converge to a certain result, no longer change, export sample card Convergence result corresponding to mouth mark-theme and theme-sample track of vehicle.The convergence result of the output is each sample card Conversion between mouth mark and the corresponding relation of theme, and the theme of sample bayonet socket mark and the theme of sample track of vehicle is closed System, above- mentioned information is obtained, the topic model, which is established, to be completed.

Further, when the track of vehicle to vehicle to be analyzed is trained, according to each sample card of above-mentioned output Conversion between mouth mark and the corresponding relation of theme, and the theme of sample bayonet socket mark and the theme of sample track of vehicle is closed System, determine the theme of every track of vehicle corresponding to vehicle to be analyzed.

Wherein, second scheduled time can be identical with first scheduled time, or different, or can have intersection.

In a kind of specific implementation, the sample, which crosses car record data, also includes corresponding vehicle by corresponding bayonet socket Second cross the car time；

Car record data is crossed according to second track division limits and the sample, it is determined that the car of corresponding vehicle Track.

It is understood that division and the division of the track of vehicle of vehicle to be analyzed for the track of vehicle of sample vehicle It is identical, it will not be described here.Wherein, obtain sample and cross in car record data the identification information for including corresponding sample vehicle.With Avoid going wrong when dividing track.

Behavioural analysis is carried out to vehicle to be analyzed using LDA models in the embodiment of the present invention, LDA models are automatically according to study The training result arrived, track of vehicle corresponding to vehicle to be analyzed is trained, theme corresponding to track of vehicle is determined, according to car Bayonet socket information corresponding to bayonet socket mark included in track, limits theme concrete meaning, further, determines car to be analyzed Behavioural information, without putting into a large amount of manpowers structure theme storehouses, effectively saved personnel time.Also, apply LDA models Behavioural analysis is carried out to vehicle to be analyzed, the interference of human factor is avoided, improves the degree of accuracy of the behavioural analysis to vehicle.

Corresponding to above method embodiment, as shown in figure 4, the embodiment of the present invention additionally provides a kind of vehicle behavioural analysis dress Put, described device can include：

Module 410 is obtained, for obtaining track of vehicle of the vehicle to be analyzed within first scheduled time, wherein, the car Track includes the bayonet socket mark for the bayonet socket that the vehicle to be analyzed is passed through；

Sort module 420, for according to default topic model, classifying to the track of vehicle of the vehicle to be analyzed；

Determining module 430, for every a kind of track of vehicle, the bayonet socket included according to every track of vehicle to identify corresponding Bayonet socket information, determine behavioural information of the vehicle to be analyzed within first scheduled time.

In a kind of specific implementation, the acquisition module 410 includes first obtains unit and the first determining unit；

The first obtains unit, for obtain the vehicle to be analyzed within first scheduled time by the of bayonet socket One crosses car record data；

In a kind of specific implementation, it is single that the sort module 420 includes input block, the second obtaining unit and division Member；

As shown in figure 5, a kind of vehicle behavioural analysis device that the embodiment of the present invention is provided can also include being used to establish The topic model of topic model establishes module, and the topic model, which establishes module, includes the 3rd obtaining unit 510, second determination list First 520, the 3rd determining unit 530 and the 4th obtaining unit 540；

3rd obtaining unit 510, for obtaining each sample car of each sample bayonet socket within second scheduled time Sample cross car record data, wherein, the sample, which crosses car record data, includes the sample identification information of sample vehicle；

Second determining unit 520, for being crossed from the sample in car record data, it is determined that the sample of each sample vehicle Sample corresponding to this identification information crosses car record data；

3rd determining unit 530, for according to default track division rule and corresponding same sample identification The sample of information crosses car record data, it is determined that the sample track of vehicle of corresponding sample vehicle, the sample track of vehicle bag Include the sample bayonet socket mark for the sample bayonet socket that vehicle is passed through；

4th obtaining unit 540, for carrying out LDA study, it is corresponding with theme to obtain each sample bayonet socket mark Transformational relation between relation, and the theme of sample bayonet socket mark and the theme of sample track of vehicle；

Second obtaining unit, the bayonet socket mark specifically for the topic model for every track of vehicle of input Know, identified according to each sample bayonet socket and determine the corresponding theme of bayonet socket mark with the corresponding relation of theme, and according to sample bayonet socket Transformational relation between the theme of mark and the theme of sample track of vehicle determines the theme of every track of vehicle, determined by output The theme of every track of vehicle；Obtain the theme of the every track of vehicle exported.

As shown in fig. 6, first determining unit includes the first sorting sub-module 610, the first determination sub-module 620, the Two determination sub-modules 630 and the 3rd determination sub-module 640；

First sorting sub-module 610, for crossing car time order and function order according to corresponding first, sequence described the One crosses car record data；

First determination sub-module 620, for according to each first cross car record data corresponding to first cross car when Between, determine that the very first time of the first adjacent car record data excessively of each two is poor；

Second determination sub-module 630, for exceeding the default first track division time when very first time difference During threshold value, the first adjacent car record data excessively of corresponding two is defined as the first track division limits, wherein, the first mistake The bayonet socket that the car time preceding first crosses bayonet socket corresponding to car record data is identified as previous track of vehicle terminal and identified, and first The bayonet socket for crossing bayonet socket corresponding to car time posterior first mistake car record data is identified as the latter track of vehicle point identification；

3rd determination sub-module 640, for crossing car note according to first track division limits and described first Data are recorded, determine the track of vehicle of the vehicle to be analyzed.

In a kind of specific implementation, the sample, which crosses car record data, also includes corresponding sample vehicle by corresponding It's the car time pasts the second of bayonet socket；

In a kind of specific implementation, the bayonet socket information includes：The longitude and latitude letter of bayonet socket corresponding to each bayonet socket mark Environment letter corresponding to architecture information or each bayonet socket mark corresponding to breath, each bayonet socket mark in bayonet socket preset range residing for bayonet socket Breath；

The determining module 430, specifically for included according to every track of vehicle bayonet socket mark corresponding to longitude and latitude Information determines region theme corresponding to per a kind of track of vehicle；According to region master corresponding to identified every a kind of track of vehicle Topic, determines the vehicle to be analyzed corresponding region theme within first scheduled time；Or, according to every track of vehicle institute Comprising bayonet socket mark corresponding to architecture information in bayonet socket preset range determine per a kind of track of vehicle corresponding to building theme； Theme is built according to corresponding to identified every a kind of track of vehicle, determines the vehicle to be analyzed in first scheduled time Building theme corresponding to interior；Or, the environmental information according to corresponding to the bayonet socket mark that every track of vehicle is included residing for bayonet socket, It is determined that environment theme corresponding to per a kind of track of vehicle；According to identified per environment theme corresponding to a kind of track of vehicle, really The fixed vehicle to be analyzed corresponding environment theme within first scheduled time.

For systems/devices embodiment, because it is substantially similar to embodiment of the method, so the comparison of description is simple Single, the relevent part can refer to the partial explaination of embodiments of method.

It should be noted that herein, such as first and second or the like relational terms are used merely to a reality Body or operation make a distinction with another entity or operation, and not necessarily require or imply and deposited between these entities or operation In any this actual relation or order.Moreover, term " comprising ", "comprising" or its any other variant are intended to Nonexcludability includes, so that process, method, article or equipment including a series of elements not only will including those Element, but also the other element including being not expressly set out, or it is this process, method, article or equipment also to include Intrinsic key element.In the absence of more restrictions, the key element limited by sentence "including a ...", it is not excluded that Other identical element also be present in process, method, article or equipment including the key element.

Can one of ordinary skill in the art will appreciate that realizing that all or part of step in above method embodiment is To instruct the hardware of correlation to complete by program, described program can be stored in computer read/write memory medium, The storage medium designated herein obtained, such as：ROM/RAM, magnetic disc, CD etc..

The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the scope of the present invention.It is all Any modification, equivalent substitution and improvements made within the spirit and principles in the present invention etc., are all contained in protection scope of the present invention It is interior.

Claims

1. a kind of vehicle behavior analysis method, it is characterised in that methods described includes：

Track of vehicle of the vehicle to be analyzed within first scheduled time is obtained, wherein, the track of vehicle includes described to be analyzed The bayonet socket mark for the bayonet socket that vehicle is passed through；

To every a kind of track of vehicle, according to bayonet socket information corresponding to the bayonet socket mark that every track of vehicle is included, it is determined that described Behavioural information of the vehicle to be analyzed within first scheduled time.

2. according to the method for claim 1, it is characterised in that described to obtain vehicle to be analyzed within first scheduled time Track of vehicle, including：

Car record data is crossed according to default track division rule and described first, determines the vehicle rail of the vehicle to be analyzed Mark.

3. according to the method for claim 1, it is characterised in that it is described according to default topic model, to the car to be analyzed Track of vehicle classified, including：

Theme identical track of vehicle is divided into one kind.

4. according to the method for claim 3, it is characterised in that the topic model is that implicit Di Li Crays are distributed LDA moulds Type；The LDA models include sample bayonet socket mark, theme and the sample track of vehicle three-decker of sample bayonet socket；

The sample for obtaining each sample vehicle of each sample bayonet socket within second scheduled time crosses car record data, wherein, institute Stating sample and crossing car record data includes the sample identification information of sample vehicle；

Car record data is crossed according to the sample of default track division rule and corresponding same sample identification information, determines institute The sample track of vehicle of corresponding sample vehicle, the sample track of vehicle include the sample card for the sample bayonet socket that vehicle is passed through Mouth mark；

Carry out LDA study, obtain each sample bayonet socket mark and the corresponding relation of theme, and the theme of sample bayonet socket mark with Transformational relation between the theme of sample track of vehicle；

The topic model identifies and theme for the bayonet socket mark of every track of vehicle of input according to each sample bayonet socket Corresponding relation determines theme corresponding to bayonet socket mark, and between the theme and the theme of sample track of vehicle identified according to sample bayonet socket Transformational relation determine the theme of every track of vehicle, the theme of every track of vehicle determined by output；

Obtain the theme of the every track of vehicle exported.

5. according to the method for claim 2, it is characterised in that described first, which crosses car record data, includes the car to be analyzed Spend the car time by the first of corresponding bayonet socket；

It is described according to default track division rule and described first crosses car record data, determine the car of the vehicle to be analyzed Track, including：

The car time is spent according to first corresponding to each first excessively car record data, determines that the first adjacent car excessively of each two records The very first time of data is poor；

When very first time difference exceedes default first track division time threshold, by corresponding two adjacent first Cross car record data and be defined as the first track division limits, wherein, the first car time preceding first excessively crossed car record data pair The bayonet socket for the bayonet socket answered is identified as previous track of vehicle terminal mark, and the first car time posterior first excessively crossed car record data The bayonet socket of corresponding bayonet socket is identified as the latter track of vehicle and plays point identification；

Car record data is crossed according to first track division limits and described first, determines the vehicle of the vehicle to be analyzed Track.

6. according to the method for claim 4, it is characterised in that the sample, which crosses car record data, also includes corresponding sample Vehicle spends the car time by the second of corresponding bayonet socket；

The sample according to default track division rule and corresponding same sample identification information crosses car record data, really The sample track of vehicle of fixed corresponding sample vehicle, including：

The second mistake car time corresponding to car record data is spent according to each sample, determines that the adjacent sample of each two crosses car record Second time difference of data；

When second time difference exceeding default second track division time threshold, by two corresponding adjacent samples Car record data is crossed as the second track division limits, wherein, second crosses car time preceding cross blocks corresponding to car record data The bayonet socket of mouth is identified as previous track of vehicle terminal mark, the second bayonet socket excessively corresponding to car time posterior mistake car record data Bayonet socket be identified as the latter track of vehicle and play point identification；

Car record data is crossed according to second track division limits and the sample, it is determined that the car of corresponding sample vehicle Track.

7. according to the method described in claim any one of 1-6, it is characterised in that the bayonet socket information includes：Each bayonet socket mark Architecture information or each bayonet socket mark pair corresponding to the latitude and longitude information of corresponding bayonet socket, each bayonet socket mark in bayonet socket preset range The environmental information residing for bayonet socket answered；

It is described to every a kind of track of vehicle, according to bayonet socket information corresponding to the bayonet socket mark that every track of vehicle is included, it is determined that Behavioural information of the vehicle to be analyzed within first scheduled time, including：

According to corresponding to latitude and longitude information corresponding to the bayonet socket mark that every track of vehicle is included determines every a kind of track of vehicle Region theme；

According to region theme corresponding to identified every a kind of track of vehicle, determine that the vehicle to be analyzed is predetermined described first Corresponding region theme in time；Or,

Determined according to the architecture information corresponding to the bayonet socket mark that every track of vehicle is included in bayonet socket preset range per a kind of Building theme corresponding to track of vehicle；

Theme is built according to corresponding to identified every a kind of track of vehicle, determines that the vehicle to be analyzed is predetermined described first Corresponding building theme in time；Or,

According to the environmental information residing for bayonet socket corresponding to the bayonet socket mark that every track of vehicle is included, it is determined that per a kind of vehicle rail Environment theme corresponding to mark；

According to environment theme corresponding to identified every a kind of track of vehicle, determine that the vehicle to be analyzed is predetermined described first Corresponding environment theme in time.

8. a kind of vehicle behavioural analysis device, it is characterised in that described device includes：

Module is obtained, for obtaining track of vehicle of the vehicle to be analyzed within first scheduled time, wherein, the track of vehicle bag Include the bayonet socket mark for the bayonet socket that the vehicle to be analyzed is passed through；

Determining module, for every a kind of track of vehicle, according to bayonet socket corresponding to the bayonet socket mark that every track of vehicle is included Information, determine behavioural information of the vehicle to be analyzed within first scheduled time.

9. device according to claim 8, it is characterised in that the acquisition module is true including first obtains unit and first Order member；

The first obtains unit, for obtaining first mistake by bayonet socket of the vehicle to be analyzed within first scheduled time Car record data, wherein, described first, which crosses car record data, includes the identification information of corresponding vehicle；

First determining unit, for crossing car record data according to default track division rule and described first, it is determined that The track of vehicle of the vehicle to be analyzed.

10. device according to claim 8, it is characterised in that the sort module includes input block, the second acquisition list Member and division unit；

11. device according to claim 10, it is characterised in that the topic model is that implicit Di Li Crays are distributed LDA Model；The LDA models include sample bayonet socket mark, theme and the sample track of vehicle three-decker of sample bayonet socket；

The topic model that described device also includes being used to establish topic model establishes module, and the topic model, which establishes module, to be included 3rd obtaining unit, the second determining unit, the 3rd determining unit and the 4th obtaining unit；

3rd obtaining unit, for obtaining the sample of each sample vehicle of each sample bayonet socket within second scheduled time Car record data is crossed, wherein, the sample, which crosses car record data, includes the sample identification information of sample vehicle；

Second determining unit, for being crossed from the sample in car record data, it is determined that the sample identification of each sample vehicle Sample corresponding to information crosses car record data；

3rd determining unit, for the sample according to default track division rule and corresponding same sample identification information This crosses car record data, it is determined that the sample track of vehicle of corresponding sample vehicle, the sample track of vehicle includes vehicle institute The sample bayonet socket mark of the sample bayonet socket of process；

4th obtaining unit, for carrying out LDA study, each sample bayonet socket mark and the corresponding relation of theme are obtained, with And the transformational relation between the theme of sample bayonet socket mark and the theme of sample track of vehicle；

Second obtaining unit, identified specifically for bayonet socket of the topic model for every track of vehicle of input, root The corresponding theme of bayonet socket mark is determined with the corresponding relation of theme according to each sample bayonet socket mark, and identified according to sample bayonet socket Transformational relation between the theme of theme and sample track of vehicle determines the theme of every track of vehicle, every car determined by output The theme of track；

Obtain the theme of the every track of vehicle exported.

12. device according to claim 9, it is characterised in that described first crosses car record data including described to be analyzed Vehicle spends the car time by the first of corresponding bayonet socket；

It is true that first determining unit includes the first sorting sub-module, the first determination sub-module, the second determination sub-module and the 3rd Stator modules；

First sorting sub-module, for crossing car time order and function order according to corresponding first, car is crossed in sequence described first Record data；

First determination sub-module, for spending the car time according to first corresponding to each first excessively car record data, it is determined that The very first time of the first adjacent car record data excessively of each two is poor；

Second determination sub-module, for when the very first time difference exceed default first track division time threshold when, The first adjacent car record data excessively of corresponding two is defined as the first track division limits, wherein, first spends the car time Preceding first bayonet socket for crossing bayonet socket corresponding to car record data is identified as previous track of vehicle terminal mark, and first when crossing car Between the posterior first bayonet socket for crossing bayonet socket corresponding to car record data be identified as the latter track of vehicle and play point identification；

3rd determination sub-module, for crossing car record data according to first track division limits and described first, Determine the track of vehicle of the vehicle to be analyzed.

13. device according to claim 11, it is characterised in that the sample, which crosses car record data, also includes corresponding sample This vehicle spends the car time by the second of corresponding bayonet socket；

It is true that 3rd determining unit includes the second sorting sub-module, the 4th determination sub-module, the 5th determination sub-module and the 6th Stator modules；

Second sorting sub-module, for crossing car time order and function order according to corresponding second, the sample that sorts crosses car Record data；

4th determination sub-module, for spending the second mistake car time corresponding to car record data according to each sample, it is determined that It's the second time difference of car record data pasts the adjacent sample of each two；

5th determination sub-module, for when second time difference exceed default second track division time threshold when, Two corresponding adjacent samples are crossed into car record data as the second track division limits, wherein, the second car time excessively existed The preceding bayonet socket for crossing bayonet socket corresponding to car record data is identified as previous track of vehicle terminal mark, and second spends the car time rear The bayonet socket for crossing bayonet socket corresponding to car record data be identified as the latter track of vehicle and play point identification；

6th determination sub-module, for crossing car record data according to second track division limits and the sample, It is determined that the track of vehicle of corresponding sample vehicle.

14. according to the device described in claim any one of 8-13, it is characterised in that the bayonet socket information includes：Each bayonet socket mark Architecture information or each bayonet socket mark corresponding to the latitude and longitude information of bayonet socket corresponding to knowledge, each bayonet socket mark in bayonet socket preset range Environmental information residing for corresponding bayonet socket；

The determining module, specifically for included according to every track of vehicle bayonet socket mark corresponding to latitude and longitude information determine Region theme corresponding to per a kind of track of vehicle；