CN112307143A - Space-time trajectory construction method, system, device and medium - Google Patents
Space-time trajectory construction method, system, device and medium Download PDFInfo
- Publication number
- CN112307143A CN112307143A CN202010870996.7A CN202010870996A CN112307143A CN 112307143 A CN112307143 A CN 112307143A CN 202010870996 A CN202010870996 A CN 202010870996A CN 112307143 A CN112307143 A CN 112307143A
- Authority
- CN
- China
- Prior art keywords
- time
- target object
- cluster
- trajectory
- position information
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000010276 construction Methods 0.000 title claims description 30
- 238000000034 method Methods 0.000 claims abstract description 38
- 238000001514 detection method Methods 0.000 claims abstract description 13
- 230000002159 abnormal effect Effects 0.000 claims description 28
- 238000004422 calculation algorithm Methods 0.000 claims description 17
- 230000008569 process Effects 0.000 claims description 6
- 238000012545 processing Methods 0.000 description 11
- 238000004458 analytical method Methods 0.000 description 10
- 238000004891 communication Methods 0.000 description 10
- 238000010586 diagram Methods 0.000 description 6
- 230000000694 effects Effects 0.000 description 6
- 230000006870 function Effects 0.000 description 4
- 238000003062 neural network model Methods 0.000 description 4
- 230000006399 behavior Effects 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 238000005259 measurement Methods 0.000 description 3
- 238000012544 monitoring process Methods 0.000 description 3
- 230000009471 action Effects 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 206010000117 Abnormal behaviour Diseases 0.000 description 1
- 206010034701 Peroneal nerve palsy Diseases 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 239000012535 impurity Substances 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000007789 sealing Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000011664 signaling Effects 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/29—Geographical information databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/906—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Probability & Statistics with Applications (AREA)
- Remote Sensing (AREA)
- Alarm Systems (AREA)
Abstract
The invention provides a method, a system, equipment and a medium for constructing a space-time trajectory, wherein the method comprises the following steps: acquiring space-time information of one or more target objects, and clustering position information of the same target object according to the space-time information to acquire a cluster; acquiring a space-time trajectory of a corresponding target object according to the clustering cluster; the invention can effectively improve the efficiency of space-time track abnormity detection.
Description
Technical Field
The invention relates to the field of artificial intelligence, in particular to a method, a system, equipment and a medium for constructing a space-time trajectory.
Background
The method is based on the analysis of space-time trajectories and the description of human behavior rules, such as footfall points, frequent removal points, abnormal points and the like, and has important significance for prevention and control work.
The traditional method relies on a service expert to perform manual analysis, and mainly judges by checking monitoring records, field visits, field surveys and other methods. The method has high requirement on experience of service experts, consumes long time, utilizes incomplete information, and is difficult to adapt to complicated and variable service scenes.
With the increase of monitoring equipment in street, community and building and the maturity of face recognition and mobile phone signaling collection technologies, the artificial intelligence technology is widely applied to behavior analysis. The face recognition algorithm can rapidly process massive snapshot data, automatically depict the space-time trajectory of each person, greatly shorten the time for manually checking monitoring records, and improve the analysis efficiency. However, the existing method still has the following defects:
first, foot-drop or frequent point aspect
1) When a person wanders or temporarily stays in a continuous acquisition device (such as a camera), a plurality of records can be generated in a short time, and the occurrence places of the actions do not calculate foothold points;
2) the foothold points, frequent points and the like of people are required to be certain areas, such as cells, supermarkets, shopping malls and the like, the areas are often provided with a plurality of exits, and the places cannot be carved only by single acquisition equipment;
3) the existing trajectory analysis mainly depends on manual setting conditions to carry out real-time query, the larger the time span is, the longer the consumed time is, so that the selected time span is usually smaller, and the comprehensiveness of the considered information is insufficient;
second, aspect of abnormal points
1) The existing abnormal track is discovered without real-time performance;
2) the considered time span is often short, the historical rule is not analyzed, if the historical rule frequently appears in the later night, the method is regarded as abnormal, and the false alarm rate is high.
3) The finding of abnormal tracks does not take into account the frequency characteristics and the regularity of human activities.
Disclosure of Invention
In view of the problems in the prior art, the invention provides a method, a system, equipment and a medium for constructing a space-time trajectory, and mainly solves the problem that the existing space-time trajectory detection depends on expert experience and is difficult to adapt to complex and variable application scenes.
In order to achieve the above and other objects, the present invention adopts the following technical solutions.
A spatiotemporal trajectory construction method comprises the following steps:
acquiring space-time information of one or more target objects, and clustering position information of the same target object according to the space-time information to acquire a cluster;
and acquiring the space-time trajectory of the corresponding target object according to the clustering cluster.
Optionally, the spatio-temporal information includes position information and occurrence time information of the target object, the occurrence frequency and the occurrence time of each position information in each cluster are obtained, the position information corresponding to the maximum occurrence frequency is used as a cluster center position of the corresponding cluster, and different cluster center positions in a predetermined time period form the spatio-temporal trajectory.
Optionally, the obtaining the occurrence frequency and the occurrence time of each piece of location information in each cluster includes: if the target object appears for multiple times within a preset time threshold, setting the appearance times of the target object as one time;
and if the distance between the positions in the cluster is smaller than a preset distance threshold, one of the positions is taken to replace the position information.
Optionally, the process of obtaining the cluster includes: and acquiring a plurality of position information of the target object and field position information adjacent to each position, and determining each cluster of the target object based on whether the density is up to reach or not according to each position and the corresponding field position.
Optionally, a time interval is set, all position information of the corresponding target object in the time interval is acquired, and the space-time trajectory of the target object in the time interval is determined.
Optionally, the method for obtaining the cluster includes one of: DBSCAN algorithm, MDCA algorithm, OPTICS algorithm.
Optionally, the method further comprises: and acquiring the identity information of the target object, and determining an abnormal point corresponding to the target object according to the space-time trajectory and the identity information.
Optionally, if the space-time trajectory corresponding to a certain target object does not conform to a preset space-time trajectory, or the space-time trajectory changes, determining that the certain target is an abnormal target, and acquiring abnormal position information in the space-time trajectory.
Optionally, before obtaining the abnormal position information in the spatiotemporal trajectory:
acquiring the number of days of occurrence of each position information of the target object in the space-time trajectory;
acquiring the occurrence time ratio of each position information in the space-time trajectory;
acquiring the time period ratio of each position information with the maximum occurrence frequency in the space-time trajectory;
and acquiring the abnormal position information according to the day number ratio, the duration ratio and the time period ratio.
A spatiotemporal trajectory construction system, comprising:
the clustering module is used for acquiring the time-space information of one or more target objects, and clustering the position information of the same target object according to the time-space information to acquire a clustering cluster; (ii) a
And the track generation module is used for acquiring the space-time track of the corresponding target object according to the clustering cluster.
Optionally, the spatio-temporal information includes position information and occurrence time information of the target object, the trajectory generation module obtains occurrence times and occurrence times of each position information in each cluster, the position information corresponding to the maximum occurrence times is used as a cluster center position of the corresponding cluster, and different cluster center positions in a predetermined time period form the spatio-temporal trajectory.
Optionally, the trajectory generation module includes a duplicate removal unit, configured to set the number of occurrences of the target object to one if the target object occurs multiple times within a preset time threshold;
and if the distance between the positions in the cluster is smaller than a preset distance threshold, one of the positions is taken to replace the position information.
Optionally, the clustering module includes a cluster acquiring unit, configured to acquire a plurality of location information of the target object and domain location information adjacent to each location, and determine, for each location and corresponding domain location, each cluster of the target object based on whether density is reachable.
Optionally, the trajectory generation module further includes a time setting unit, configured to set a time interval, acquire all position information of the target object corresponding to the time interval, and determine a spatiotemporal trajectory of the target object within the time interval.
Optionally, the system comprises an anomaly detection module, configured to obtain identity information of a target object, and determine an anomaly point corresponding to the target object according to the spatio-temporal trajectory and the identity information; the abnormity detection module is connected with the track generation module.
An apparatus, comprising:
one or more processors; and
one or more machine readable media having instructions stored thereon that, when executed by the one or more processors, cause the apparatus to perform the spatiotemporal trajectory construction method.
One or more machine-readable media having instructions stored thereon, which when executed by one or more processors, cause an apparatus to perform the spatiotemporal trajectory construction method.
As described above, the method, system, device and medium for constructing spatiotemporal trajectories according to the present invention have the following advantageous effects.
The space-time trajectories of a plurality of target objects can be detected simultaneously, and the detection efficiency is improved; the space-time trajectory is constructed through clustering, and target object behavior analysis is carried out based on the space-time trajectory, so that the detection accuracy can be effectively improved.
Drawings
FIG. 1 is a flow chart of a spatiotemporal trajectory construction method according to an embodiment of the present invention.
FIG. 2 is a block diagram of a spatiotemporal trajectory construction system in an embodiment of the present invention.
Fig. 3 is a schematic structural diagram of a terminal device in an embodiment of the present invention.
Fig. 4 is a schematic structural diagram of a terminal device in an embodiment of the present invention.
FIG. 5 is a comparison graph of density clustering effects according to an embodiment of the present invention.
FIG. 6 is a diagram illustrating an abnormal point according to an embodiment of the present invention.
Detailed Description
The embodiments of the present invention are described below with reference to specific embodiments, and other advantages and effects of the present invention will be easily understood by those skilled in the art from the disclosure of the present specification. The invention is capable of other and different embodiments and of being practiced or of being carried out in various ways, and its several details are capable of modification in various respects, all without departing from the spirit and scope of the present invention. It is to be noted that the features in the following embodiments and examples may be combined with each other without conflict.
It should be noted that the drawings provided in the following embodiments are only for illustrating the basic idea of the present invention, and the components related to the present invention are only shown in the drawings rather than drawn according to the number, shape and size of the components in actual implementation, and the type, quantity and proportion of the components in actual implementation may be changed freely, and the layout of the components may be more complicated.
Referring to FIG. 1, the present invention provides a method for constructing spatiotemporal trajectories, which includes steps S01-S02.
In step S01, spatio-temporal information of one or more target objects is obtained, and location information of the same target object is clustered according to the spatio-temporal information, obtaining a cluster:
in one embodiment, the spatiotemporal information includes a target object occurrence time, a target object occurrence location, and the like. If the video image is acquired through a camera at the entrance of the supermarket, the time stamp of the video image can be acquired at the same time, and the time of the target object in the video image can be analyzed according to the time stamp. Furthermore, the frame image corresponding to the video can be divided into a plurality of grids according to a certain proportion, the longitude and latitude of the grids are analyzed, the grid position is generated according to the longitude and latitude, and the target object appearance position is directly obtained according to the grid area corresponding to the target object. The grid may be a rectangular grid, for example, the corresponding frame image may be divided into 3X3 grids, and the position of the target object appearing in the grid is determined according to the longitude and latitude of the grid.
In an embodiment, a time interval may be set according to the occurrence time of the target object, and position information of the target object in the images acquired by the plurality of image acquisition devices in the time interval may be acquired. If the value is obtained by analyzing the video timestamp, the appearance time of the target object in a certain frame of image is 2000 year 2 and 14 days, the time interval can be set to be one week (2 month and 10 days in 2000 year-2 month and 16 days in 2000 year), and all the activity position information of the target object collected in the one week time is obtained.
In one embodiment, a plurality of position information of the target object and domain position information adjacent to each position are obtained, and for each position and corresponding domain position, each cluster of the target object is determined based on whether the density is reachable. In particular, a density clustering algorithm may be employed to cluster the location information. For example, a mall a includes four doors, namely an east door, a west door, a south door and a north door, each door is provided with a camera device, and positions of the same target object in images acquired by the four doors can be clustered into the same cluster in a preset time interval. In an embodiment, the position information of a plurality of target objects in a time interval may be clustered simultaneously. The Clustering method can adopt one of Density Clustering algorithms such as DBSCAN algorithm (sensitivity-Based Spatial Clustering of Application with Noise), MDCA algorithm (Maximum sensitivity Clustering Application), OPTICS algorithm (Ordering Points To Identify the Clustering Structure), and the like.
Taking the DBSCAN algorithm as an example, a class in the algorithm represents a cluster of samples with reachable density, the similarity measurement is defined as reachable density, and all samples with reachable density are found out through the DBSCAN.
Assuming a sample set with position information, it can be expressed as (x)1,x2,...,xm) In DBSCAN, in order to describe the relationship of sample distribution, the following concepts are defined:
1) α -neighborhood: for xjE.g. D, whose alpha-neighborhood contains the sum x in the sample set DiIs not more than
A subsample set of, i.e.
The number of subsamples is recorded as | Nα|。
2) Core object: for any sample xje.D if N corresponds to its alpha-neighborhoodα(xj) At least containing MinPts samples, i.e. if Nα(xj) Greater than or equal to MinPts, then xjIs the core object.
3) The density is up to: if xiAt xjIn the alpha-neighborhood of (a), and xjIs a core object, then called xiFrom xjThe density is up to. The converse is not necessarily true unless and xiIs also a core object.
4) The density can reach: if xiFrom xjDensity is direct, and xjFrom xkDensity through, then xiFrom xkThe density can be reached. The density can meet the sealing property.
The density reachable is similarity measurement, all points in the cluster and core objects in the cluster are reachable in average density due to the closeness of the density reachable, otherwise, the density reachable is not a cluster, so that the density reachable can cluster samples, wherein parameters related to the density reachable include
MinPts and distance metric distance (x)i,xj)。
5) Noise points: noise points are defined for non-core points and points that cannot be reached by the density of core points.
DBSCAN algorithm flow:
inputting: sample set (x) of location information1,x2,...,xm) Setting neighborhood parameters
The sample distance measurement mode can adopt Euclidean distance, Manhattan distance and the like.
And (3) outputting: cluster division C ═ C1,c2,...,ck)
Arbitrarily taking a point x from the sample setjCalculating all the slaves xjRegarding distance (x)i,xj) And point x where the MinPts density can be reachedi。
If xjIf the object is a core object, finding a cluster;
if xjIs a boundary point, not from xjThe point with the reachable density is marked as an isolated point, and the next point in the sample set is accessed;
the above process is repeated until all points in the sample set have been processed. Clustering can be performed to obtain clusters and isolated points. Referring to fig. 5, after clustering, black dots in the graph represent cluster clusters, and gray dots represent isolated dots.
Labels may be set for representing clusters and outliers, such as (labels ═ 0,1,2,3, …) and an outlier label ═ 1.
Through the steps, the clustering cluster label corresponding to the position information in each grid can be obtained. Geographically close locations can be grouped into the same cluster by DBSCAN.
In step S02, a spatiotemporal trajectory corresponding to the target object is obtained according to the cluster.
In one embodiment, the location information with all cluster labels greater than-1 may be taken as one sample set. The number of occurrences of each position information in the sample set is obtained. If the acquired images are analyzed, the situation that the target object appears at the door of the supermarket for 5 times in one week and appears at the door of the building A for 10 times is known. The sequential method counts the number of occurrences of location information in each cluster.
Furthermore, the position information with the largest occurrence frequency of the target object in each cluster can be obtained as the cluster center position of the corresponding cluster according to the descending order of the occurrence frequency of the position information, and the space-time trajectory of the target object is formed by different cluster center positions in a preset time period. If the clustering cluster with label equal to 1 contains three position information of a, b and c, the position a appears 5 times, the position b appears 2 times, and the position c appears 1 time, then a is the position information with the largest number of appearance, and a is taken as the cluster center position.
In an embodiment, a time threshold and a distance threshold may be set to deduplicate clusters. If the target object appears for multiple times within a preset time threshold, setting the appearance times of the target object as one time; if the time threshold is set to be 10 minutes and the target object appears for multiple times, only one position information record is kept, and the appearance times of the target object in the time period are recorded as one time. The distance threshold is set to be 20 meters, the distances among a, b and c are all within the distance threshold, and only one of the three positions a, b and c is reserved on the side to replace the three position information. In this way, redundant position information resulting from the target object wandering or lingering near a position may be screened out.
After the repetition, the clustering cluster and the set of all the isolated points can be spliced together to form the space-time trajectory of the target object in the time interval.
In an embodiment, identity information of the target object, such as identity card information and face information of the monitored object, may be obtained, and the abnormal point of the target object may be obtained according to time information of each position point in the spatio-temporal trajectory. Specifically, the time information may be encoded according to information such as date, time type, etc. corresponding to the time information, and different time types may be distinguished by the encoding. The time types may include working days, holidays, working hours, early morning hours, and the like. The time interval may be divided into a plurality of time types according to a specific target object. If the target object is a working group, working days are Monday to Friday according to work and rest time of the working group, and if the target object appears at the gate of an office building on a sunday, the sunday is considered as an abnormal time period; the conventional time period is from 8:00 in the morning to 6 o' clock in the evening; 1 in the morning: 00 to 5:00 is an abnormal period. The encoding rules may be set to (e.g., workday encoding of 1, holiday encoding of 5, regular time period encoding of 1-2, abnormal time period encoding of 8-16).
In one embodiment, according to the divided time types, acquiring the number of days of occurrence of each piece of position information of a target object in the space-time trajectory; acquiring the occurrence time ratio of each position information in the space-time trajectory; acquiring the time period ratio of each position information with the maximum occurrence frequency in the space-time trajectory; and acquiring the abnormal position information according to the day number ratio, the duration ratio and the time period ratio. If the space-time trajectory corresponding to a certain target object does not conform to the preset space-time trajectory or the space-time trajectory changes, determining that the certain target is an abnormal target, and acquiring abnormal position information in the space-time trajectory, wherein if the number of days that the target object should appear in the building A in a week is 5 days but the target object only appears for 3 days, working days appear in the position B for two days, the position B is an isolated point, the position B is an abnormal position, and the corresponding two days are time abnormal points.
In one embodiment, the input information includes the appearance time and the appearance position of the target object, and the derived features can be obtained according to the input information. Specifically, cluster clusters and isolated points obtained from the historical spatio-temporal information of the target object in step S01, and information such as time type codes and occurrence times can be extracted.
The clustering center of each cluster can be obtained by adopting a K-means algorithm. Firstly, randomly selecting K position information as an initial clustering center. The distance between each location information and the respective initial cluster center is then calculated, and each location information is assigned to the cluster center closest to it. The cluster centers and the position information assigned to them represent a cluster. Once all of the location information has been assigned, the cluster center for each cluster is recalculated based on the location information existing in the cluster. This process will be repeated until some termination condition is met. The termination condition may be any one of the following:
1) no (or a minimum number of) location information is reassigned to different clusters.
2) No (or minimal) cluster centers change again.
3) The sum of squared errors reaches a local minimum.
After the clustering centers are obtained, the distance between the position information in the input information and each clustering center is calculated. Further, the nearest neighbor position information of the position information in all the cluster clusters and the isolated points and the input information is obtained, and the distance between the position information in the input information and the nearest neighbor position information is calculated. The relevance of the input information and the historical spatiotemporal information can be obtained through distance calculation. And constructing a derivative feature according to the calculated distance information.
In one embodiment, the time type in the input information can be encoded by using WOE encoding, and the time type is adjusted by using the WOE encoding result. If the working time of the target object changes, the working time changes from daytime working to night working, the sensitive time period also changes, and the adjustment is needed according to the actual situation. WOE is a form of encoding of the original arguments, whose formula is calculated as follows:
wherein, Bi is the number of abnormal time points corresponding to the derivative characteristic i, B is the total number of the abnormal time points, Gi is the number of normal time points corresponding to the derivative characteristic i, and G is the total number of the normal time points.
If the value of the WOE code is in the set range, the historical time type does not need to be adjusted, and if the value of the WOE code exceeds the set range, the time information in the input information corresponds to a large number of abnormal points, so that misjudgment is easy to occur, and the division rule of the time type is adjusted again.
In one embodiment, the derived features may be feature selected using a GBDT model. The GBDT model is a Boosting integration model, which can be expressed as:
wherein T (x; theta)m) The weak classifiers are generated in each iteration, a CART decision tree is selected, and gini indexes are used for measuring the impurity degree or uncertainty of data.
The importance of the feature X at the node m, i.e. the variation of gini coefficients before and after the node m branches.
VIMj＝GIm-GIl-GIr
Wherein, GIlAnd GIrRespectively representing gini coefficients, GI, of two new nodes after branchingmRepresenting gini coefficients before branching. Assuming that the GBDT has a total of n trees, the importance scores of the features after normalization are:
wherein, VIMjIs the sum of the importance of the features j in the n trees, Σ VIMiIs the sum of the importance of all features in n trees.
And selecting the derivative features with higher scores according to the importance scores.
And inputting the selected features into a pre-trained deep neural network model, and predicting whether the input information is abnormal information. The method can be used for training the deep neural network model after the position information in the historical clustering result is selected through the GBDT characteristics. Referring to fig. 6, the black dot set represents outliers.
The input information can be real-time stream data, the real-time stream data can be divided into a plurality of batches of data by Spark-streaming, and the real-time data is input into the deep neural network model in batches to be processed.
Referring to fig. 2, the present embodiment provides a spatiotemporal trajectory construction system for implementing the spatiotemporal trajectory construction method described in the foregoing method embodiments. Since the technical principle of the system embodiment is similar to that of the method embodiment, repeated description of the same technical details is omitted.
In an embodiment, the spatiotemporal trajectory construction system includes a clustering module 10 and a trajectory generation module 11, the clustering module 10 is configured to assist in performing step S01 described in the foregoing method embodiment, and the detection module 11 is configured to perform step S02 described in the foregoing method embodiment.
Optionally, the spatio-temporal information includes position information and occurrence time information of the target object, the trajectory generation module obtains occurrence times and occurrence time of each position information in each cluster, the position information corresponding to the maximum occurrence times is used as a cluster center position of the corresponding cluster, and different cluster center positions in a predetermined time period form the spatio-temporal trajectory.
Optionally, the trajectory generation module includes a duplicate removal unit, configured to set the number of occurrences of the target object to one if the target object occurs multiple times within a preset time threshold;
and if the distance between the positions in the cluster is smaller than a preset distance threshold, one of the positions is taken to replace the information of the positions.
Optionally, the trajectory generation module further includes a time setting unit, configured to set a time interval, acquire all position information of the target object within the time interval, and determine a spatiotemporal trajectory of the target object within the time interval.
Optionally, the system comprises an anomaly detection module, a time-space trajectory determination module and a time-space trajectory determination module, wherein the anomaly detection module is used for acquiring identity information of a target object and determining an anomaly point corresponding to the target object according to the time-space trajectory and the identity information; and the abnormity detection module is connected with the track generation module.
An embodiment of the present application further provides an apparatus, which may include: one or more processors; and one or more machine readable media having instructions stored thereon that, when executed by the one or more processors, cause the apparatus to perform the method of fig. 1. In practical applications, the device may be used as a terminal device, and may also be used as a server, where examples of the terminal device may include: the mobile terminal includes a smart phone, a tablet computer, an electronic book reader, an MP3 (Moving Picture Experts Group Audio Layer III) player, an MP4 (Moving Picture Experts Group Audio Layer IV) player, a laptop, a vehicle-mounted computer, a desktop computer, a set-top box, an intelligent television, a wearable device, and the like.
The embodiment of the present application also provides a non-volatile readable storage medium, where one or more modules (programs) are stored in the storage medium, and when the one or more modules are applied to a device, the device may be caused to execute instructions (instructions) of steps included in the spatiotemporal trajectory construction method in fig. 1 according to the embodiment of the present application.
Fig. 3 is a schematic diagram of a hardware structure of a terminal device according to an embodiment of the present application. As shown, the terminal device may include: an input device 1100, a first processor 1101, an output device 1102, a first memory 1103, and at least one communication bus 1104. The communication bus 1104 is used to implement communication connections between the elements. The first memory 1103 may include a high-speed RAM memory, and may also include a non-volatile storage NVM, such as at least one disk memory, and the first memory 1103 may store various programs for performing various processing functions and implementing the method steps of the present embodiment.
Alternatively, the first processor 1101 may be, for example, a Central Processing Unit (CPU), an Application Specific Integrated Circuit (ASIC), a Digital Signal Processor (DSP), a Digital Signal Processing Device (DSPD), a Programmable Logic Device (PLD), a Field Programmable Gate Array (FPGA), a controller, a microcontroller, a microprocessor, or other electronic components, and the processor 1101 is coupled to the input device 1100 and the output device 1102 through a wired or wireless connection.
Optionally, the input device 1100 may include a variety of input devices, such as at least one of a user-oriented user interface, a device-oriented device interface, a software programmable interface, a camera, and a sensor. Optionally, the device interface facing the device may be a wired interface for data transmission between devices, or may be a hardware plug-in interface (e.g., a USB interface, a serial port, etc.) for data transmission between devices; optionally, the user-facing user interface may be, for example, a user-facing control key, a voice input device for receiving voice input, and a touch sensing device (e.g., a touch screen with a touch sensing function, a touch pad, etc.) for receiving user touch input; optionally, the programmable interface of the software may be, for example, an entry for a user to edit or modify a program, such as an input pin interface or an input interface of a chip; the output devices 1102 may include output devices such as a display, audio, and the like.
In this embodiment, the processor of the terminal device includes a function for executing each module of the speech recognition apparatus in each device, and specific functions and technical effects may refer to the above embodiments, which are not described herein again.
Fig. 4 is a schematic hardware structure diagram of a terminal device according to another embodiment of the present application. Fig. 4 is a specific embodiment of fig. 3 in an implementation process. As shown, the terminal device of the present embodiment may include a second processor 1201 and a second memory 1202.
The second processor 1201 executes the computer program code stored in the second memory 1202 to implement the method described in fig. 1 in the above embodiment.
The second memory 1202 is configured to store various types of data to support operations at the terminal device. Examples of such data include instructions for any application or method operating on the terminal device, such as messages, pictures, videos, and so forth. The second memory 1202 may include a Random Access Memory (RAM) and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.
Optionally, the first processor 1201 is provided in the processing assembly 1200. The terminal device may further include: communication component 1203, power component 1204, multimedia component 1205, speech component 1206, input/output interfaces 1207, and/or sensor component 1208. The specific components included in the terminal device are set according to actual requirements, which is not limited in this embodiment.
The processing component 1200 generally controls the overall operation of the terminal device. The processing assembly 1200 may include one or more second processors 1201 to execute instructions to perform all or part of the steps of the method illustrated in fig. 1 described above. Further, the processing component 1200 can include one or more modules that facilitate interaction between the processing component 1200 and other components. For example, the processing component 1200 can include a multimedia module to facilitate interaction between the multimedia component 1205 and the processing component 1200.
The power supply component 1204 provides power to the various components of the terminal device. The power components 1204 may include a power management system, one or more power sources, and other components associated with generating, managing, and distributing power for the terminal device.
The multimedia components 1205 include a display screen that provides an output interface between the terminal device and the user. In some embodiments, the display screen may include a Liquid Crystal Display (LCD) and a Touch Panel (TP). If the display screen includes a touch panel, the display screen may be implemented as a touch screen to receive an input signal from a user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundary of a touch or slide action, but also detect the duration and pressure associated with the touch or slide operation.
The voice component 1206 is configured to output and/or input voice signals. For example, the voice component 1206 includes a Microphone (MIC) configured to receive external voice signals when the terminal device is in an operational mode, such as a voice recognition mode. The received speech signal may further be stored in the second memory 1202 or transmitted via the communication component 1203. In some embodiments, the speech component 1206 further comprises a speaker for outputting speech signals.
The input/output interface 1207 provides an interface between the processing component 1200 and peripheral interface modules, which may be click wheels, buttons, etc. These buttons may include, but are not limited to: a volume button, a start button, and a lock button.
The sensor component 1208 includes one or more sensors for providing various aspects of status assessment for the terminal device. For example, the sensor component 1208 may detect an open/closed state of the terminal device, relative positioning of the components, presence or absence of user contact with the terminal device. The sensor assembly 1208 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact, including detecting the distance between the user and the terminal device. In some embodiments, the sensor assembly 1208 may also include a camera or the like.
The communication component 1203 is configured to facilitate communications between the terminal device and other devices in a wired or wireless manner. The terminal device may access a wireless network based on a communication standard, such as WiFi, 2G or 3G, or a combination thereof. In one embodiment, the terminal device may include a SIM card slot therein for inserting a SIM card therein, so that the terminal device may log onto a GPRS network to establish communication with the server via the internet.
As can be seen from the above, the communication component 1203, the voice component 1206, the input/output interface 1207 and the sensor component 1208 referred to in the embodiment of fig. 4 can be implemented as the input device in the embodiment of fig. 3.
In conclusion, the spatiotemporal trajectory construction method, the spatiotemporal trajectory construction system, the spatiotemporal trajectory construction equipment and the spatiotemporal trajectory construction medium can prolong the clustering analysis time interval to several months, store the analysis result in an off-line calculation mode, generate the trajectory portrait in an off-line mode, perform efficient batch query when analyzing foothold points, and improve the analysis efficiency; the problem of data sparseness is solved to a certain extent by expanding a time interval, adjusting a minimum unit of time analysis (a time threshold for duplicate removal), and a gridding and clustering method; compared with a mining method based on rules, the method has stronger generalization capability by using a machine learning algorithm and is more suitable for complex and changeable scenes; by using the flow calculation and the neural network model, the abnormal behavior can be detected in real time; and (4) passing. Therefore, the invention effectively overcomes various defects in the prior art and has high industrial utilization value.
The foregoing embodiments are merely illustrative of the principles and utilities of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or change the above-mentioned embodiments without departing from the spirit and scope of the present invention. Accordingly, it is intended that all equivalent modifications or changes which can be made by those skilled in the art without departing from the spirit and technical spirit of the present invention be covered by the claims of the present invention.
Claims (17)
1. A spatiotemporal trajectory construction method is characterized by comprising the following steps:
acquiring space-time information of one or more target objects, and clustering position information of the same target object according to the space-time information to acquire a cluster;
and acquiring the space-time trajectory of the corresponding target object according to the clustering cluster.
2. The spatiotemporal trajectory construction method according to claim 1, wherein the spatiotemporal information includes position information and occurrence time information of a target object, the occurrence frequency and the occurrence time of each position information in each cluster are obtained, the position information corresponding to the maximum occurrence frequency is used as a cluster center position of the corresponding cluster, and different cluster center positions within a predetermined time period constitute the spatiotemporal trajectory.
3. The spatiotemporal trajectory construction method according to claim 2,
the acquiring the occurrence frequency and the occurrence time of each position information in each cluster includes: if the target object appears for multiple times within a preset time threshold, setting the appearance times of the target object as one time;
and if the distance between the positions in the cluster is smaller than a preset distance threshold, one of the positions is taken to replace the position information.
4. The spatio-temporal trajectory construction method according to claim 1, wherein the process of obtaining cluster clusters comprises: and acquiring a plurality of position information of the target object and field position information adjacent to each position, and determining each cluster of the target object based on whether the density is up to reach or not according to each position and the corresponding field position.
5. The spatiotemporal trajectory construction method according to claim 1, wherein a time interval is set, all position information of the corresponding target object in the time interval is acquired, and the spatiotemporal trajectory of the target object in the time interval is determined.
6. The spatio-temporal trajectory construction method according to any one of claims 1 to 5, wherein the method of obtaining the cluster comprises one of: DBSCAN algorithm, MDCA algorithm, OPTICS algorithm.
7. The spatiotemporal trajectory construction method according to claim 1, further comprising: and acquiring the identity information of the target object, and determining an abnormal point corresponding to the target object according to the space-time trajectory and the identity information.
8. The spatiotemporal trajectory construction method according to claim 1 or 7, wherein if the spatiotemporal trajectory corresponding to a certain target object does not conform to a preset spatiotemporal trajectory or the spatiotemporal trajectory changes, it is determined that the certain target is an abnormal target, and abnormal position information in the spatiotemporal trajectory is obtained.
9. The spatiotemporal trajectory anomaly detection method according to claim 8, characterized in that before acquiring anomaly location information in the spatiotemporal trajectory:
acquiring the number of days of occurrence of each position information of the target object in the space-time trajectory;
acquiring the occurrence time ratio of each position information in the space-time trajectory;
acquiring the time period ratio of each position information with the maximum occurrence frequency in the space-time trajectory;
and acquiring the abnormal position information according to the day number ratio, the duration ratio and the time period ratio.
10. A spatiotemporal trajectory construction system, comprising:
the clustering module is used for acquiring the time-space information of one or more target objects, and clustering the position information of the same target object according to the time-space information to acquire a clustering cluster; (ii) a
And the track generation module is used for acquiring the space-time track of the corresponding target object according to the clustering cluster.
11. The spatiotemporal trajectory construction system according to claim 10, wherein the spatiotemporal information includes position information and occurrence time information of a target object, the trajectory generation module obtains the occurrence number and the occurrence time of each position information in each cluster, the position information corresponding to the maximum occurrence number is used as a cluster center position of the corresponding cluster, and different cluster center positions within a predetermined time period constitute the spatiotemporal trajectory.
12. The spatiotemporal trajectory construction system according to claim 11, wherein the trajectory generation module includes a deduplication unit configured to set the number of occurrences of the target object to one if the target object occurs multiple times within a preset time threshold;
and if the distance between the positions in the cluster is smaller than a preset distance threshold, one of the positions is taken to replace the position information.
13. The spatiotemporal trajectory construction system according to claim 10, wherein the clustering module includes a cluster acquisition unit configured to acquire a plurality of location information of the target object and domain location information adjacent to each location, and determine each cluster of the target object based on whether density is reachable for each location and corresponding domain location.
14. The spatiotemporal trajectory construction system according to claim 10, wherein the trajectory generation module further comprises a time setting unit for setting a time interval, acquiring all position information of the corresponding target object within the time interval, and determining the spatiotemporal trajectory of the target object within the time interval.
15. The spatiotemporal trajectory construction system according to claim 10, comprising an anomaly detection module for obtaining identity information of a target object and determining an anomaly point corresponding to the target object according to the spatiotemporal trajectory and the identity information; the abnormity detection module is connected with the track generation module.
16. An apparatus, comprising:
one or more processors; and
one or more machine-readable media having instructions stored thereon that, when executed by the one or more processors, cause the apparatus to perform the method recited by one or more of claims 1-9.
17. One or more machine-readable media having instructions stored thereon, which when executed by one or more processors, cause an apparatus to perform the method recited by one or more of claims 1-9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010870996.7A CN112307143A (en) | 2020-08-26 | 2020-08-26 | Space-time trajectory construction method, system, device and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010870996.7A CN112307143A (en) | 2020-08-26 | 2020-08-26 | Space-time trajectory construction method, system, device and medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112307143A true CN112307143A (en) | 2021-02-02 |
Family
ID=74483794
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010870996.7A Pending CN112307143A (en) | 2020-08-26 | 2020-08-26 | Space-time trajectory construction method, system, device and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112307143A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112925948A (en) * | 2021-02-05 | 2021-06-08 | 上海依图网络科技有限公司 | Video processing method and device, medium, chip and electronic equipment thereof |
| CN113283653A (en) * | 2021-05-27 | 2021-08-20 | 大连海事大学 | Ship track prediction method based on machine learning and AIS data |
| CN116049464A (en) * | 2022-08-05 | 2023-05-02 | 荣耀终端有限公司 | Image sorting method and electronic equipment |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103246706A (en) * | 2013-04-09 | 2013-08-14 | 哈尔滨工程大学 | Method of clustering motion trajectories of vehicle objects in road network space |
| CN105894358A (en) * | 2016-03-31 | 2016-08-24 | 百度在线网络技术（北京）有限公司 | Commuting order identification method and device |
| CN107133269A (en) * | 2017-04-01 | 2017-09-05 | 中国人民解放军国防科学技术大学 | Frequent location track generation method and device based on mobile target |
| CN110324787A (en) * | 2019-06-06 | 2019-10-11 | 东南大学 | A kind of duty residence acquisition methods of mobile phone signaling data |
| CN111459997A (en) * | 2020-03-16 | 2020-07-28 | 中国科学院计算技术研究所 | Frequent mode increment mining method of space-time trajectory data and electronic equipment |
-
2020
- 2020-08-26 CN CN202010870996.7A patent/CN112307143A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103246706A (en) * | 2013-04-09 | 2013-08-14 | 哈尔滨工程大学 | Method of clustering motion trajectories of vehicle objects in road network space |
| CN105894358A (en) * | 2016-03-31 | 2016-08-24 | 百度在线网络技术（北京）有限公司 | Commuting order identification method and device |
| CN107133269A (en) * | 2017-04-01 | 2017-09-05 | 中国人民解放军国防科学技术大学 | Frequent location track generation method and device based on mobile target |
| CN110324787A (en) * | 2019-06-06 | 2019-10-11 | 东南大学 | A kind of duty residence acquisition methods of mobile phone signaling data |
| CN111459997A (en) * | 2020-03-16 | 2020-07-28 | 中国科学院计算技术研究所 | Frequent mode increment mining method of space-time trajectory data and electronic equipment |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112925948A (en) * | 2021-02-05 | 2021-06-08 | 上海依图网络科技有限公司 | Video processing method and device, medium, chip and electronic equipment thereof |
| CN113283653A (en) * | 2021-05-27 | 2021-08-20 | 大连海事大学 | Ship track prediction method based on machine learning and AIS data |
| CN113283653B (en) * | 2021-05-27 | 2024-03-26 | 大连海事大学 | Ship track prediction method based on machine learning and AIS data |
| CN116049464A (en) * | 2022-08-05 | 2023-05-02 | 荣耀终端有限公司 | Image sorting method and electronic equipment |
| CN116049464B (en) * | 2022-08-05 | 2023-10-20 | 荣耀终端有限公司 | Image sorting method and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110032670B (en) | Method, device and equipment for detecting abnormity of time sequence data and storage medium | |
| CN112307143A (en) | Space-time trajectory construction method, system, device and medium | |
| Huang et al. | Modeling and visualizing regular human mobility patterns with uncertainty: An example using Twitter data | |
| US11232685B1 (en) | Security system with dual-mode event video and still image recording | |
| CN113052079B (en) | Regional passenger flow statistical method, system, equipment and medium based on face clustering | |
| CN112100305B (en) | Trajectory determination method, device, equipment and medium based on multi-source data | |
| CN109558416A (en) | A kind of detection method traded extremely, device and storage medium | |
| Hossain et al. | Modeling and assessing quality of information in multisensor multimedia monitoring systems | |
| US10742939B1 (en) | Security video data processing systems and methods | |
| CN111881000A (en) | Fault prediction method, device, equipment and machine readable medium | |
| Vries et al. | Application of machine learning techniques to predict anomalies in water supply networks | |
| CN113947800A (en) | Face confidence method, system, equipment and medium based on space-time collision | |
| Appice et al. | Data mining techniques in sensor networks: Summarization, interpolation and surveillance | |
| CN101790717B (en) | Machine vision system for enterprise management | |
| CN112580474B (en) | Target object detection method, system, device and medium based on computer vision | |
| CN111178455B (en) | Image clustering method, system, device and medium | |
| US9460393B2 (en) | Inference of anomalous behavior of members of cohorts and associate actors related to the anomalous behavior based on divergent movement from the cohort context centroid | |
| US10914811B1 (en) | Locating a source of a sound using microphones and radio frequency communication | |
| CN111091152A (en) | Image clustering method, system, device and machine readable medium | |
| CN114913600A (en) | Electricity stealing detection method and device for electricity meter, electronic equipment and storage medium | |
| Hiriyannaiah et al. | Data Reduction Techniques in Fog Data Analytics for IoT Applications | |
| CN111694979A (en) | Archive management method, system, equipment and medium based on image | |
| Mohan et al. | Modelling large scale camera networks for identification and tracking: an abstract framework | |
| CN110689375A (en) | Information interaction method, system, equipment and medium based on portable equipment | |
| Kang et al. | Outlier Behavior Detection for Indoor Environment Based on t-SNE Clustering. |
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 |