CN110941278B - Dynamic safety analysis method in station - Google Patents
Dynamic safety analysis method in station Download PDFInfo
- Publication number
- CN110941278B CN110941278B CN201911328001.8A CN201911328001A CN110941278B CN 110941278 B CN110941278 B CN 110941278B CN 201911328001 A CN201911328001 A CN 201911328001A CN 110941278 B CN110941278 B CN 110941278B
- Authority
- CN
- China
- Prior art keywords
- track
- preset
- point
- motion
- station
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000004458 analytical method Methods 0.000 title claims abstract description 46
- 238000000034 method Methods 0.000 claims abstract description 26
- 238000007635 classification algorithm Methods 0.000 claims abstract description 14
- 238000004422 calculation algorithm Methods 0.000 claims description 48
- 238000012549 training Methods 0.000 claims description 39
- 238000001914 filtration Methods 0.000 claims description 18
- 230000002159 abnormal effect Effects 0.000 claims description 13
- 238000007781 pre-processing Methods 0.000 claims description 9
- 230000002123 temporal effect Effects 0.000 claims description 7
- 238000012216 screening Methods 0.000 claims description 3
- 238000005516 engineering process Methods 0.000 abstract description 6
- 239000011159 matrix material Substances 0.000 description 14
- 230000008569 process Effects 0.000 description 7
- 238000012546 transfer Methods 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 4
- 238000011161 development Methods 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000000739 chaotic effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000003064 k means clustering Methods 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 231100001263 laboratory chemical safety summary Toxicity 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010187 selection method Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0231—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means
- G05D1/0242—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means using non-visible light signals, e.g. IR or UV signals
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Electromagnetism (AREA)
- Aviation & Aerospace Engineering (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Alarm Systems (AREA)
Abstract
The embodiment of the invention provides an in-station dynamic security analysis method. The method comprises the steps of acquiring the motion trail of each passenger in a station by adopting a location-based service LBS; classifying the motion tracks according to track models of preset types by adopting a preset classification algorithm, and counting the track number corresponding to each track model; if the track number of a certain track model exceeds a preset alarm threshold, sending alarm information according to statistical information of motion tracks belonging to the certain track model, acquiring the motion tracks of all passengers through an LBS technology to judge the corresponding track model, and sending the alarm information when the track number of the track model exceeds the alarm threshold, so that alarm prediction can be carried out on safety risks in a station more accurately.
Description
Technical Field
The invention relates to the technical field of rail transit, in particular to an in-station dynamic security analysis method.
Background
Along with the rapid development of subway construction in China, the complexity of subway lines is also rapidly improved, and a large number of subway stations for multi-line (3 lines and more) transfer, such as three-line transfer stations of Beijing's Sichuan gate station, song Guzhuang station, shenzhen Futian station and the like, and four-line transfer stations of Shanghai century large-scale road station, long Yanglu station and the like, are further appeared. In a multi-line transfer station, because stations where different lines are located are different, the cross of transfer lines in the station is caused, and under the impact of large passenger flow in the peak, holidays and major activities in the morning and evening, the safety operation of the subway is greatly risked.
In order to cope with the risks, a method of manual work and a camera safety analysis mode are adopted at present, the number or distribution density of passenger groups in the station is detected through a camera, and then on-site guiding and commanding are carried out by staff in the station. However, the installation position of the camera is affected by the space in the station, so that a large amount of shielding is caused, the detection range is limited, and a large amount of cameras are arranged, so that complicated wiring, installation and debugging work is brought. Whereas the process of manual supervision can only analyze the people stream density in a single area.
Therefore, the existing security analysis method cannot accurately predict the security risk in the station.
Disclosure of Invention
Because the existing method has the problems, the embodiment of the invention provides an in-station dynamic security analysis method.
In a first aspect, an embodiment of the present invention provides an in-station dynamic security analysis method, which is characterized by including:
acquiring the motion trail of each passenger in the station by adopting a location-based service LBS;
classifying the motion tracks according to track models of preset types by adopting a preset classification algorithm, and counting the track number corresponding to each track model;
and if the track number of a certain track model exceeds a preset alarm threshold, sending alarm information according to the statistical information of the motion track belonging to the certain track model.
Further, if the number of tracks of a certain track model exceeds a preset alarm threshold, sending alarm information according to statistical information of motion tracks belonging to the certain track model, including:
if the track number of a certain track model exceeds the alarm threshold, executing a preset track point clustering algorithm on each motion track belonging to the certain track model to obtain a region of interest (ROI) and the track point number corresponding to each ROI;
according to the analysis of the number of the track points corresponding to each ROI, alarm information is sent; the alarm information comprises an alarm reason and an alarm area.
Further, the step of executing a preset track point clustering algorithm on each motion track belonging to the certain track model to obtain a region of interest ROI and the number of track points corresponding to each ROI specifically includes:
obtaining an entering point, an exiting point and a stopping point in each motion trail according to a preset interest point analysis method; the entering point is the first track point of the motion track, the exiting point is the last track point of the motion track, and the stopping point is the track point obtained by screening according to a preset CB-SMoT algorithm;
respectively carrying out a preset DBSCAN algorithm on an entry point, an exit point and a stop point of each motion trail to obtain ROIs and corresponding relations between each ROI and each entry point, each exit point and each stop point; the ROI comprises at least one entrance region, at least one exit region and at least one stop region;
and counting to obtain the number of track points corresponding to each ROI according to the corresponding relation between each ROI and each entering point, each leaving point and each stopping point.
Further, after classifying each motion trail, the in-station dynamic security analysis method further includes:
counting the motion trail of the trail model which does not belong to the preset category to obtain the number of abnormal trail;
and judging the degree of confusion in the station according to the number of the abnormal tracks.
Further, after the motion trail of each passenger in the station is obtained by adopting the location-based service LBS, the method for dynamic security analysis in the station further comprises:
performing preset data preprocessing on the motion trail, and eliminating the motion trail which does not meet preset filtering conditions; wherein, the preset filtering conditions comprise:
the spatial distance and the temporal distance between the entering point and the leaving point of the motion trail exceed a preset spatial threshold value and a preset temporal threshold value;
the number of track points contained in the motion track exceeds a preset number threshold.
Further, after sending the alarm information, the in-station dynamic security analysis method further includes:
and according to the alarm information, obtaining a suggested passing path.
Further, the in-station dynamic security analysis method further comprises the following steps:
and according to a pre-stored first training set containing the historical motion trail, adopting the preset trail point clustering algorithm to obtain the trail model of the preset type.
Further, the obtaining the track model of the preset type by adopting the preset track point clustering algorithm according to a pre-stored first training set including the historical motion track specifically includes:
acquiring the first training set, performing the preset data preprocessing on each historical motion trail, and removing the historical motion trail which does not meet the preset filtering condition to obtain a second training set;
executing the preset track point clustering algorithm on each historical motion track in the second training set to obtain an ROI corresponding to the second training set;
according to the ROI corresponding to the second training set, removing the historical motion trail which does not meet the preset complete trail condition from the second training set to obtain a third training set; wherein, the preset complete track condition comprises: the entry point and the exit point of the historical motion trail respectively belong to an entry area and an exit area;
executing a preset k-meas clustering algorithm on each historical motion track in the third training set to obtain k track types, and calculating prototype motion tracks corresponding to each track type;
combining the track types meeting the preset combining conditions by adopting a preset combining algorithm according to the prototype motion track of each track type to obtain a track model of the preset type; wherein the number of the preset categories is less than or equal to k.
Further, the classification algorithm is specifically a K nearest neighbor KNN classification algorithm.
Further, the preset merging algorithm is specifically a connected subgraph search of an undirected graph.
According to the in-station dynamic safety analysis method provided by the embodiment of the invention, the motion trail of each passenger is obtained through the LBS technology to judge the corresponding trail model, and when the trail number of the trail models exceeds the alarm threshold, alarm information is sent, so that the in-station safety risk can be more accurately predicted in an alarm way.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a method for dynamic security analysis in a station according to an embodiment of the present invention;
FIG. 2 is a flow chart of another in-station dynamic security analysis method according to an embodiment of the present invention;
FIG. 3 is a flow chart of a method of dynamic security analysis in a further station according to an embodiment of the present invention;
FIG. 4 is a flow chart of a dynamic security analysis method in another station according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Fig. 1 is a flowchart of a method for dynamic security analysis in a station according to an embodiment of the present invention, as shown in fig. 1, where the method includes:
and step S01, acquiring the motion trail of each passenger in the station by adopting a location-based service LBS.
Location based services (Location Based Services, LBS) technology is to acquire location information of a user terminal through a mobile communication network.
The security analysis system of the embodiment of the invention acquires the position information of the user terminal of each passenger in the station through the LBS technology, and sorts the position information on a map as the track points of the passenger, thereby obtaining the motion track P= { P of each passenger in the period from entering the station to leaving the station 0 ,p 1 ,…,p n And realizing the full-flow tracking of each single target in the station. Each locus point p i The corresponding position information at least comprises acquisition Time Time i And three-dimensional space coordinates (x i ,y i ,z i ). The motion trail is a three-dimensional space coordinate sequence arranged according to the acquisition time.
Step S02, classifying the motion trajectories according to a trajectory model of a preset type by adopting a preset classification algorithm, and counting the number of trajectories corresponding to each trajectory model.
The safety analysis system acquires a preset type of track model in advance, wherein each track model comprises a corresponding typical track, and the typical track is also composed of a three-dimensional space coordinate sequence.
And matching the motion trail of the passenger with the typical trail of each trail model through a preset classification algorithm. And determining a track model to which the motion track of each passenger belongs according to the matching result.
Further, the classification algorithm is specifically a K nearest neighbor KNN classification algorithm.
There are many classification algorithms that can be used to match motion trajectories to typical trajectories, and only the K nearest neighbor (K-NearestNeighbor, KNN) classification algorithm is given here as an example for illustration, the K value takes 1. The distances between the motion track and each typical track are calculated respectively, and the adopted distance measurement can be set according to actual needs, for example, euclidean distance can be adopted. And taking a track model corresponding to a typical track with the nearest motion track as the type of the motion track. When the Euclidean distance between two tracks is calculated, if the number of track points contained in the two tracks is different, the shorter track needs to be complemented to the same number by adopting the numerical value of the last track point.
In order to increase the calculation speed when calculating the euclidean distance, the three-dimensional space coordinate may be first reduced in size to a two-dimensional space coordinate by projecting the three-dimensional space coordinate in the direction of the Z bearing, and then the euclidean distance may be calculated.
After classifying each motion trail, counting the trail number of the motion trail corresponding to each trail model.
Step S03, if the track number of a certain track model exceeds a preset alarm threshold, sending alarm information according to the statistical information of the motion track belonging to the certain track model.
If the number of tracks counted by a certain track model exceeds a preset alarm threshold value in a preset period, a large number of passengers are judged to move in the station in a short period according to the typical track corresponding to the track model, so that safety risks exist in the area where the typical track corresponding to the track model passes, and corresponding alarm information needs to be sent to related personnel.
The specific content of the alarm information can be determined by statistical information obtained by further statistical analysis of all motion trajectories belonging to the trajectory model.
According to the embodiment of the invention, the motion trail of each passenger is obtained through the LBS technology to judge the corresponding trail model, and the alarm information is sent when the trail number of the trail models exceeds the alarm threshold value, so that the alarm prediction on the safety risk in the station can be more accurately carried out.
Fig. 2 is a flowchart of another in-station dynamic security analysis method according to an embodiment of the present invention, as shown in fig. 2, where the step S03 specifically includes:
step S031, if the track number of a certain track model exceeds the alarm threshold, executing a preset track point clustering algorithm on each motion track belonging to the certain track model to obtain a region of interest (ROI) and the track point number corresponding to each ROI.
When the track number of a certain track model in the current period exceeds the alarm threshold value, all the motion tracks which belong to the track model and are counted in the current period can be further analyzed.
Counting the track points contained in all the motion tracks belonging to the track model, and executing a preset track point clustering algorithm to obtain a plurality of regions of interest (Region of Interest, ROI), wherein the ROI is a region with concentrated track points. And simultaneously, counting the corresponding track points of each ROI to obtain the track point number of each ROI.
Step S032, according to the number analysis of the track points corresponding to each ROI, sending alarm information; the alarm information comprises an alarm reason and an alarm area.
The number of track points corresponding to each ROI and the position of the station corresponding to each ROI are analyzed, so that the alarm reason of the current alarm, specific alarm areas and other contents are determined, and the contents are recorded into alarm information for transmission. Taking as an example the ROIs corresponding to the entrance a and the exit B, an example is illustrated:
if the number of the track points of the ROI corresponding to the entrance A is more and the number of the track points of the ROI corresponding to the exit B is also more, the condition that a large number of passengers enter from the entrance A and a large number of passengers exit from the exit B is indicated, and the alarm is because the subway station is in a point-to-point running mode, and the entrance A and the exit B need to be subjected to current limiting and dredging;
if the number of the track points of the ROI corresponding to the entrance A is large and the number of the track points of the ROI corresponding to the exit B is small, the alarm is caused by the fact that the subway station is in an operation mode of centralized entrance of a certain entrance, and at the moment, the entrance A needs to be subjected to current limiting and dredging;
if the number of the track points of the ROI corresponding to the entrance A is smaller and the number of the track points of the ROI corresponding to the exit B is also larger, the alarm is caused by the fact that the vicinity of the exit B of the subway station is in a burst passenger flow state, and at the moment, the exit B needs to be subjected to current limiting and dredging;
and if the number of the track points of the ROI corresponding to the entrance A is smaller and the number of the track points of the ROI corresponding to the exit B is smaller, the fact that the alarm reason does not exist is indicated, and the subway station is in the operation valley.
According to the embodiment of the invention, the ROI and the corresponding track point number are obtained by executing the track point clustering algorithm on all the motion tracks belonging to a certain track model, so that the alarm information comprising the subway running mode and the alarm area is further analyzed and obtained, and the safety risk is predicted more accurately.
Based on the above embodiment, further, the step S031 specifically includes:
step S0311, obtaining an entry point, an exit point and a stopping point in each motion trail according to a preset interest point analysis method; the entering point is the first track point of the motion track, the exiting point is the last track point of the motion track, and the stopping point is the track point obtained by screening according to a preset CB-SMoT algorithm.
The specific calculation process of the track point clustering algorithm can be set according to actual needs, and only one of the specific calculation processes is illustrated in the embodiment of the invention.
The method comprises the steps of firstly obtaining interesting points of each motion track, wherein the interesting points can be divided into an entering point, an exiting point and a stopping point.
The method for acquiring the entry point and the exit point is simple, and the first track point in all track points contained in the motion track can be directly used as the entry point of the motion track, and the last track point can be used as the exit point of the motion track.
The stopping point reflects the stay time of the passenger nearby the track point, and the efficiency condition of the passenger passing through the track point can be reflected to a certain extent.
The stopping point can be specifically obtained by adopting a track stopping point identification algorithm CB-SMoT algorithm based on a correlation coefficient. By calculating the ith trace point p i The average passing speed of the track points with the preset number in front and back is smaller than a preset speed threshold value, if the average passing speed is smaller than the preset speed threshold value, the track point p i I.e. can be referred to as a core point, which is the stopping point. For example, if the preset number is 3, the average passing speed is:
said (x) i-1 ，y i-1 ，z i-1 ) Sum Time i-1 Is the locus point p i-1 And acquisition time, the (x) i+1 ，y i+1 ，z i+1 ) Sum Time i+1 Is the locus point p i+1 Three-dimensional space coordinates and acquisition time of (a).
The speed threshold may be set according to a preset normal human walking speed, for example, set to 30% of the normal human walking speed, 0.3m/s.
Step S0312, respectively performing a preset DBSCAN algorithm on the entry point, the exit point and the stop point of each motion trail to obtain ROIs and the corresponding relation between each ROI and each entry point, each exit point and each stop point; the ROI includes at least one entrance region, at least one exit region, and at least one stop region.
And then, respectively performing a preset Density-based clustering algorithm (Density-Based Spatial Clustering of Applications with Noise, DBSCAN) on the entry points, the exit points and the stop points corresponding to all the motion trajectories, so as to obtain the region of interest ROI. Obtaining at least one access area by performing DBSCAN on all access points; performing DBSCAN on all the leaving points to obtain at least one leaving area; and performing DBSCAN on all stopping points to obtain at least one stopping area.
The DBSCAN clustering algorithm may find clusters in the data with noisy points. Specifically, the DBSCAN algorithm needs to give two parameters in advance, one is the sample point neighborhood radius Eps, and the other is the minimum point number MinPts in the neighborhood. For the locus point p i In other words, if the number of points of the trajectory Eps (i) of points in its Eps neighborhood>MinPts, trace point p i Referred to as core points. If Eps (i)<MinPts, and trace point p i Not in the Eps neighborhood of any other core point, then trace point p i Is defined as the noise point. For core point p i And p is as follows j If the two track points are in the neighborhood of each other, the two core points and the track points in the neighborhood form clusters, the clustering relation is transferred along with the neighborhood retrieval, the areas formed by the clusters are gradually expanded, and finally, the areas formed by the clusters are ROIs, and the corresponding relation between the ROIs and the entry points, the exit points and the stopping points is determined by labeling the corresponding ROI labels on the track points in the areas.
The domain radius Eps and the minimum number of points in the neighborhood MinPts may be set according to the size of a specific station.
Step S0313, counting to obtain the number of track points corresponding to each ROI according to the corresponding relation between each ROI and each entry point, each exit point and each stop point.
And counting the track points with the same ROI label to obtain the track point number corresponding to each ROI.
According to the embodiment of the invention, the entering point, the leaving point and the stopping point of each motion track are obtained, and the ROI and the track point number corresponding to each ROI are obtained through the DBSCAN algorithm and used for obtaining the subsequent alarm information, so that the safety risk can be predicted more accurately.
Based on the above embodiment, further, after classifying each motion trail, the in-station dynamic security analysis method further includes:
counting the motion trail of the trail model which does not belong to the preset category to obtain the number of abnormal trail;
and judging the degree of confusion in the station according to the number of the abnormal tracks.
In the process of classifying the acquired motion trail through a preset classification algorithm, if a trail model matched with the motion trail is not found, classifying the motion trail into an abnormal trail for indicating that phenomena such as retrograde, traversing or fighting occur. And counting the number of the abnormal tracks. Therefore, the chaotic program passing in the station can be judged according to the abnormal track number, and a corresponding alarm is given.
According to the embodiment of the invention, the motion track which does not belong to any track model is judged to be the abnormal track, and the number of the abnormal tracks is counted, so that the disorder degree of the traffic in the station can be rapidly determined according to the number of the abnormal tracks, and the safety risk alarm can be timely carried out.
Fig. 3 is a flowchart of a further method for in-station dynamic security analysis according to an embodiment of the present invention, as shown in fig. 3, where the method for in-station dynamic security analysis after step S01 further includes:
step S011, performing preset data preprocessing on the motion trail, and eliminating the motion trail which does not meet preset filtering conditions; wherein, the preset filtering conditions comprise:
the spatial distance and the temporal distance between the entering point and the leaving point of the motion trail exceed a preset spatial threshold value and a preset temporal threshold value; the entry point is the first track point of the motion track, and the exit point is the last track point of the motion track;
the number of track points contained in the motion track exceeds a preset number threshold.
In the actual application process, when the motion trail of each passenger is obtained according to the LBS technology, some special abnormal trails are received, for example, the motion trail caused by signal interruption or mistransmission, so that the motion trail can be subjected to data preprocessing through preset filtering conditions, and the motion trail which does not meet the filtering conditions can be removed.
The filtering conditions may be set according to actual needs, and only some examples of the filtering conditions are given in the embodiment of the present invention.
In the process of acquiring the motion trail P= { P 0 ,p 1 ,…,p n After (a) comparing the first trace point p 0 And the last trace point p n And obtaining the space distance and the time distance between the two track points, and comparing the space distance and the time distance with a preset space threshold value and a preset time threshold value.
If the space distance between the two track points is smaller than the space threshold value, the motion track is too short, the filtering condition is not met, and the motion track needs to be removed;
if the time distance between the two track points is smaller than the time threshold, the time of the motion track is too short, and the filtering condition is not met and needs to be removed.
In addition, the number n+1 of the track points included in the motion track can be counted and compared with a preset number threshold, and if the number of the track points is smaller than the number threshold, it is determined that the track points included in the motion track are too few and do not meet the filtering condition, and the track points need to be removed.
The spatial threshold, temporal threshold, and number threshold may be set according to the size and scale of the station.
According to the embodiment of the invention, the obtained motion trail is subjected to data preprocessing, so that the motion trail which does not meet the preset filtering condition is removed, and the judgment and alarm of safety risk are accelerated.
Based on the above embodiment, further, the in-station dynamic security analysis method further includes:
and according to the alarm information, obtaining a suggested passing path.
As can be seen from the above embodiments, when the security analysis system determines that there is a security risk, corresponding alarm information is sent out to give out a specific alarm reason and an alarm area. Thus, when a new passenger enters the station, a suggested traffic path may be sent to the passenger for avoiding the current alert zone.
According to the embodiment of the invention, the recommended passing path is obtained for the passengers according to the alarm information, so that the passengers can pass through the platform more quickly, and the further improvement of the safety risk is avoided.
Fig. 4 is a flowchart of a dynamic security analysis method in another station according to an embodiment of the present invention, as shown in fig. 4, where the dynamic security analysis method in a station further includes:
and S00, obtaining the track model of the preset type by adopting the preset track point clustering algorithm according to a pre-stored first training set containing the historical motion track.
As can be seen from the above embodiments, the track model of the preset type for classification is acquired before classifying the acquired motion track. Specifically, the method can be obtained by performing a preset track point clustering algorithm on a pre-stored first training set consisting of a large number of historical motion tracks.
In the practical application process, the track model is different along with the change of time or surrounding environment due to the development of transportation. Thus, the trajectory model may be updated periodically.
Further, the step S00 specifically includes:
step S001, acquiring the first training set, executing the preset data preprocessing on each historical motion trail, and eliminating the historical motion trail which does not meet the preset filtering condition to obtain a second training set.
Step S002, executing the preset track point clustering algorithm on each historical motion track in the second training set to obtain the ROI corresponding to the second training set;
step S003, according to the ROI corresponding to the second training set, removing the historical motion trail which does not meet the preset complete trail condition from the second training set to obtain a third training set; wherein, the preset complete track condition comprises: the entry point and the exit point of the historical motion trail respectively belong to an entry area and an exit area.
In order to ensure the integrity of the obtained trajectory model as much as possible, it is therefore necessary to first clean and screen each historical motion trajectory. And according to the preset data preprocessing, removing the historical motion trail which does not meet the preset filtering condition from the first training set to obtain the first training set.
And performing a preset track point clustering algorithm on all the historical motion tracks in the second training set to obtain the ROI corresponding to the second training set, wherein the ROI at least comprises an entering area and an exiting area.
And according to the corresponding relation between the entry point and the exit point of each historical motion track and the entry area and the exit area, if the entry point of the historical motion track does not belong to any entry area or the exit point does not belong to any exit area, judging that the historical motion track does not meet the preset complete track condition, and removing the historical motion track from the second training set. And combining the remaining historical motion trajectories with the full trajectory condition into a third training set.
And S004, executing a preset k-meas clustering algorithm on each historical motion track in the third training set to obtain k track types, and calculating a prototype motion track corresponding to each track type.
And K track types can be obtained by executing a preset K-means clustering algorithm on each historical motion track in the third training set. The specific method is as follows:
the number of track points contained in all the historical motion tracks in the first training set is complemented to the same length. Obtaining the similarity w between the historical motion tracks by calculating the Euclidean distance between the historical motion tracks ij 。
Of course, the similarity between the historical motion trajectories may also be performed by the dynamic time warping algorithm DTW or the longest common subsequence algorithm LCSS, which is not specifically limited herein.
Constructing an adjacency matrix W= { W according to the similarity between the historical motion tracks ij And then calculate the laplace matrix:
L sym ＝I-D -1/2 WD -1/2 ，
wherein the Laplacian period matrix L sym Has the following properties:
matrix L sym Is symmetrical and positive;
matrix L sym Is 9, the corresponding feature vector is
The feature vector is a column vector.
Matrix L sym There are M non-negative real eigenvalues, 0=λ 1 ≤λ 2 …≤λ M 。
To Laplace matrix L sym And (3) carrying out eigenvalue decomposition, arranging the eigenvalues from small to large, and selecting the first k eigenvalues and the eigenvectors corresponding to the first k eigenvalues to form an Mxk matrix U. The rows of the matrix U can be regarded as new characteristic representation after the original data are transformed, k categories can be obtained by finally clustering the rows of the matrix U by using k-means, the number of the rows of the matrix U is equal to the total number of the historical motion trajectories, and therefore the number of the categories of the rows of the matrix U is equal to the number of the trajectory types of the historical motion trajectories.
The k value in the k-means algorithm is the same number as the number of columns k of the U matrix, and the selection method of k is the same as the common rule of the k-means algorithm, and a relatively large value can be selected.
At this time, each historical motion track in the third training set is allocated to a corresponding track type, a historical motion track corresponding to each track type is extracted, contained track points are averaged, and a prototype motion track of the track type is obtained through calculation.
Step S005, combining the track types meeting the preset combining conditions by adopting a preset combining algorithm according to the prototype motion track of each track type to obtain a track model of the preset type; wherein the number of the preset categories is less than or equal to k.
Because the k value is larger, the track types meeting the preset merging conditions can be merged by adopting a preset merging algorithm according to the prototype motion track corresponding to each track type, so that the number of classifications is reduced.
Further, the preset merging algorithm is specifically a connected subgraph search of an undirected graph.
There are many preset merging algorithms, and only one of them is illustrated in the embodiment of the present invention.
The merging algorithm is abstracted to solve the problem of connected subgraph search of an undirected graph: according to the prototype motion track of each track type, calculating a similarity matrix S= { S ij Let each prototype motion track be a node V of graph g= (V, E) i If v of two clustering prototypes i And v j Similarity s of (2) ij Above a certain threshold eta e And if the nodes are considered to be connected with the nodes, the corresponding track types can be combined. After merging, a track model of a preset type can be obtained.
According to the embodiment of the invention, the preset track point clustering algorithm is executed on the prestored historical motion tracks to obtain the track model of the preset type, and the track model is used for classifying the motion tracks of all passengers, so that the safety risk can be predicted more quickly and accurately.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (9)
1. A method of dynamic security analysis in a station, comprising:
acquiring the motion trail of each passenger in the station by adopting a location-based service LBS;
classifying the motion tracks according to track models of preset types by adopting a preset classification algorithm, and counting the track number corresponding to each track model;
if the track number of a certain track model exceeds a preset alarm threshold, sending alarm information according to the statistical information of the motion track belonging to the certain track model;
if the number of tracks of a certain track model exceeds a preset alarm threshold, sending alarm information according to statistical information of motion tracks belonging to the certain track model, wherein the method specifically comprises the following steps:
if the track number of a certain track model exceeds the alarm threshold, executing a preset track point clustering algorithm on each motion track belonging to the certain track model to obtain a region of interest (ROI) and the track point number corresponding to each ROI;
according to the analysis of the number of the track points corresponding to each ROI, sending the alarm information; the alarm information comprises an alarm reason and an alarm area.
2. The method for in-station dynamic security analysis according to claim 1, wherein the performing a preset trajectory point clustering algorithm on each motion trajectory belonging to the certain trajectory model to obtain a region of interest ROI and the number of trajectory points corresponding to each ROI specifically includes:
obtaining an entering point, an exiting point and a stopping point in each motion trail according to a preset interest point analysis method; the entering point is the first track point of the motion track, the exiting point is the last track point of the motion track, and the stopping point is the track point obtained by screening according to a preset CB-SMoT algorithm;
respectively carrying out a preset DBSCAN algorithm on an entry point, an exit point and a stop point of each motion trail to obtain each ROI and the corresponding relation between each ROI and each entry point, each exit point and each stop point; the ROI comprises at least one entrance region, at least one exit region and at least one stop region;
and counting to obtain the number of track points corresponding to each ROI according to the corresponding relation between each ROI and each entering point, each leaving point and each stopping point.
3. The in-station dynamic security analysis method according to claim 2, wherein after classifying each motion trajectory, the in-station dynamic security analysis method further comprises:
counting the motion trail of the trail model which does not belong to the preset category to obtain the number of abnormal trail;
and judging the degree of confusion in the station according to the number of the abnormal tracks.
4. The in-station dynamic security analysis method according to claim 3, wherein after the acquiring of the motion trail of each passenger in the station using the location-based service LBS, the in-station dynamic security analysis method further comprises:
performing preset data preprocessing on the motion trail, and eliminating the motion trail which does not meet preset filtering conditions; wherein, the preset filtering conditions comprise:
the spatial distance and the temporal distance between the entering point and the leaving point of the motion trail exceed a preset spatial threshold value and a preset temporal threshold value;
the number of track points contained in the motion track exceeds a preset number threshold.
5. The in-station dynamic security analysis method according to claim 4, wherein after transmitting the alarm information, the in-station dynamic security analysis method further comprises:
and according to the alarm information, obtaining a suggested passing path.
6. The in-station dynamic security analysis method according to claim 5, further comprising:
and according to a pre-stored first training set containing the historical motion trail, adopting the preset trail point clustering algorithm to obtain the trail model of the preset type.
7. The method for in-station dynamic security analysis according to claim 6, wherein the obtaining the track model of the preset type by using the preset track point clustering algorithm according to the pre-stored first training set including the historical motion track specifically includes:
acquiring the first training set, performing the preset data preprocessing on each historical motion trail, and removing the historical motion trail which does not meet the preset filtering condition to obtain a second training set;
executing the preset track point clustering algorithm on each historical motion track in the second training set to obtain an ROI corresponding to the second training set;
according to the ROI corresponding to the second training set, removing the historical motion trail which does not meet the preset complete trail condition from the second training set to obtain a third training set; wherein, the preset complete track condition comprises: the entry point and the exit point of the historical motion trail respectively belong to an entry area and an exit area;
executing a preset k-meas clustering algorithm on each historical motion track in the third training set to obtain k track types, and calculating prototype motion tracks corresponding to each track type;
combining the track types meeting the preset combining conditions by adopting a preset combining algorithm according to the prototype motion track of each track type to obtain a track model of the preset type; wherein the number of the preset categories is less than or equal to k.
8. The in-station dynamic security analysis method according to claim 7, wherein the classification algorithm is in particular a K nearest neighbor KNN classification algorithm.
9. The method for dynamic security analysis in a station according to claim 7, wherein the preset merging algorithm is specifically a connected subgraph search of an undirected graph.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911328001.8A CN110941278B (en) | 2019-12-20 | 2019-12-20 | Dynamic safety analysis method in station |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911328001.8A CN110941278B (en) | 2019-12-20 | 2019-12-20 | Dynamic safety analysis method in station |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110941278A CN110941278A (en) | 2020-03-31 |
| CN110941278B true CN110941278B (en) | 2023-05-23 |
Family
ID=69912647
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911328001.8A Active CN110941278B (en) | 2019-12-20 | 2019-12-20 | Dynamic safety analysis method in station |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110941278B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111523577A (en) * | 2020-04-13 | 2020-08-11 | 南京烽火星空通信发展有限公司 | Mass trajectory similarity calculation method based on improved LCSS algorithm |
| CN112268563B (en) * | 2020-10-23 | 2022-03-04 | 重庆越致科技有限公司 | Method and device for detecting traffic state and track of pedestrian on track |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106529815A (en) * | 2016-11-15 | 2017-03-22 | 同济大学 | Estimation method of passenger trip spatial-temporal trajectory of urban rail transit network and application thereof |
| CN109445304A (en) * | 2018-09-25 | 2019-03-08 | 交控科技股份有限公司 | A kind of intelligent fault analysis system and method based on cab signal |
Family Cites Families (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8229458B2 (en) * | 2007-04-08 | 2012-07-24 | Enhanced Geographic Llc | Systems and methods to determine the name of a location visited by a user of a wireless device |
| CN103327091A (en) * | 2013-06-17 | 2013-09-25 | 北京交通大学 | System and method for obtaining passenger track and behavioral parameter |
| CN104036139B (en) * | 2014-06-12 | 2017-07-07 | 中国科学院软件研究所 | A kind of mobile object track monitoring method |
| CN104657424A (en) * | 2015-01-21 | 2015-05-27 | 段炼 | Clustering method for interest point tracks under multiple temporal and spatial characteristic fusion |
| CN106454716B (en) * | 2015-08-13 | 2019-11-19 | 株式会社理光 | Object analysis method, object analysis device and object analysis system |
| CN109923595B (en) * | 2016-12-30 | 2021-07-13 | 同济大学 | An urban road traffic anomaly detection method based on floating car data |
| CN107798557B (en) * | 2017-09-30 | 2020-09-25 | 平安科技（深圳）有限公司 | Electronic device, service place recommendation method based on LBS data and storage medium |
| CN109784970B (en) * | 2018-12-13 | 2020-09-25 | 交控科技股份有限公司 | Service recommendation method and device based on AFC passenger riding data |
| CN109829119B (en) * | 2018-12-26 | 2023-04-28 | 成都熊谷油气科技有限公司 | Information early warning method based on LBS big data in intelligent network |
| CN109961106B (en) * | 2019-04-18 | 2022-03-18 | 北京百度网讯科技有限公司 | Training method and device for trajectory classification model, and electronic device |
| CN110111566B (en) * | 2019-04-19 | 2021-07-06 | 腾讯科技（深圳）有限公司 | Trajectory prediction method, apparatus and storage medium |
| CN110309383B (en) * | 2019-06-17 | 2021-07-13 | 武汉科技大学 | Clustering Analysis Method of Ship Trajectory Based on Improved DBSCAN Algorithm |
| CN110517495A (en) * | 2019-09-05 | 2019-11-29 | 四川东方网力科技有限公司 | Confirmation method, device, equipment and the storage medium of track of vehicle classification |
| CN110570658B (en) * | 2019-10-23 | 2022-02-01 | 江苏智通交通科技有限公司 | Method for identifying and analyzing abnormal vehicle track at intersection based on hierarchical clustering |
-
2019
- 2019-12-20 CN CN201911328001.8A patent/CN110941278B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106529815A (en) * | 2016-11-15 | 2017-03-22 | 同济大学 | Estimation method of passenger trip spatial-temporal trajectory of urban rail transit network and application thereof |
| CN109445304A (en) * | 2018-09-25 | 2019-03-08 | 交控科技股份有限公司 | A kind of intelligent fault analysis system and method based on cab signal |
Non-Patent Citations (1)
| Title |
|---|
| 贾若然 ; 刘曙光 ; 孙启龙 ; .基于位置轨迹数据的用户相似性分析.计算机与数字工程.2016,第44卷(第08期),第1523-1527页. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110941278A (en) | 2020-03-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Saunier et al. | Probabilistic collision prediction for vision-based automated road safety analysis | |
| CN113011367B (en) | An Abnormal Behavior Analysis Method Based on Target Trajectory | |
| Park et al. | Real-time prediction of secondary incident occurrences using vehicle probe data | |
| CN111462488A (en) | An Intersection Safety Risk Assessment Method Based on Deep Convolutional Neural Network and Intersection Behavior Feature Model | |
| CN111860282A (en) | Method and system for passenger flow statistics of subway section and pedestrian retrograde detection | |
| CN103986910A (en) | A method and system for counting passenger flow based on intelligent analysis camera | |
| Abdullah et al. | Semantic segmentation based crowd tracking and anomaly detection via neuro-fuzzy classifier in smart surveillance system | |
| Salmane et al. | A video-analysis-based railway–road safety system for detecting hazard situations at level crossings | |
| Giannakeris et al. | Speed estimation and abnormality detection from surveillance cameras | |
| CN114038056B (en) | A method for identifying jumping and squatting ticket evasion behaviors | |
| CN108052859A (en) | A kind of anomaly detection method, system and device based on cluster Optical-flow Feature | |
| CN110941278B (en) | Dynamic safety analysis method in station | |
| CN114612860A (en) | Identification and prediction method of passenger flow in rail transit station based on computer vision | |
| CN103646253A (en) | Bus passenger flow statistics method based on multi-motion passenger behavior analysis | |
| CN114926422B (en) | Method and system for detecting passenger flow of getting on and off vehicles | |
| CN115423735A (en) | Passenger flow volume statistical method and system | |
| CN111723664A (en) | Pedestrian counting method and system for open type area | |
| Lumentut et al. | Evaluation of recursive background subtraction algorithms for real-time passenger counting at bus rapid transit system | |
| CN114782857A (en) | Method and system for monitoring abnormal behavior of passenger in railway passenger station | |
| CN117649428A (en) | Suspicious person tracking method, device, equipment and medium based on cloud edge cooperation | |
| Chen et al. | Comogcn: Coherent motion aware trajectory prediction with graph representation | |
| Saunier et al. | Clustering vehicle trajectories with hidden Markov models application to automated traffic safety analysis | |
| CN112382068A (en) | Station waiting line crossing detection system based on BIM and DNN | |
| Kaviani et al. | A new method for traffic density estimation based on topic model | |
| KR20210037925A (en) | Passenger counting apparatus using computer vision and passenger monitoring system thereof |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |