in the formula, R (i, j) represents a second correlation index between a preset person i and a preset person j, the value range of R (i, j) is 0-1, RC (i, j) represents the correlation times between the preset person i and the preset person j, and the preset person i is different from the preset person j. RC (resistor-capacitor) capacitor_maxRepresenting a preset threshold of association times.

The association frequency between the first preset personnel and the second preset personnel is the sum of the co-occurrence frequency of each area, and the co-occurrence of one area is that the first preset personnel and the second preset personnel appear in the area within the preset time difference.

Optionally, thecontrol server 10 may obtain the sensing data of the first preset person and the sensing data of the second preset person, which are obtained in a first preset time period (the first preset time period is a historical time period, and the time duration of the first preset time period may be preset), and divide the obtained sensing data into a plurality of groups according to the collection location. Subsequently, for each set of perception data, thecontrol server 10 determines the number of perception data sets satisfying the condition, and determines the determined number as the number of co-occurrence times of the area. Here, the sensing data set satisfying the condition may be: the acquisition time difference is smaller than the preset time difference, and the two pieces of sensing data which are different from the preset personnel are acquired (namely, one piece of sensing data corresponds to the first preset personnel, and the other piece of sensing data corresponds to the second preset personnel). Then, thecontrol server 10 accumulates the co-occurrence times of the respective areas, where the accumulated co-occurrence times are the association times between the first preset person and the second preset person. Thecontrol server 10 may store the number of associations.

Illustratively, as shown in fig. 6, the sensing data acquired by thecontrol server 10 in the first preset time period includes sensing data 1 (corresponding to person 1), sensing data 2 (corresponding to person 2), sensing data 3 (corresponding to person 2), sensing data 4 (corresponding to person 1), and sensing data 5 (corresponding to person 1). Thecontrol server 10 can determine that the acquisition place is an area a and the acquisition time istime 1 according to thesensing data 1; according to the perception data 2, the acquisition place is determined to be an area a, and the acquisition time is determined to be time 2; according to the perception data 3, the acquisition place is determined to be an area c, and the acquisition time is determined to be time 3; according to the perception data 4, the acquisition place is determined to be an area a, and the acquisition time is determined to be 4; the acquisition place is the area c and the acquisition time is the time 5 can be determined according to the perception data 5. After the perception data 1-5 are acquired, thecontrol server 10 divides the acquired perception data according to the acquisition place to obtain a second perception data group a and a second perception data group b. The second sensing data group a includessensing data 1, sensing data 2, and sensing data 4, and the second sensing data group b includes sensing data 3 and sensing data 5. If the time difference between thetime 1 and the time 2 and the time difference between the time 2 and the time 4 are smaller than the preset time difference, thecontrol server 10 determines that the number of co-occurrences of theperson 1 and the person 2 in the area a is 2 (since thesensing data 1 and the sensing data 4 are both sensing data of theperson 1, the time difference between the acquisition time of thesensing data 1 and the acquisition time of the sensing data 4 may not be considered in the process of calculating the number of co-occurrences). If the time difference between time 3 and time 5 is smaller than the preset time difference, thecontrol server 10 determines that the number of times of co-occurrence ofperson 1 and person 2 in the area c is 1. In this way, thecontrol server 10 determines that the number of associations betweenperson 1 and person 2 is 3 (i.e., 2+1 — 3).

After determining the second association index between every two preset persons appearing in the target area within the preset time period, thecontrol server 10 may calculate the first association index according to all the determined second association indexes.

Alternatively, thecontrol server 10 may calculate a mean value of all the determined second correlation indexes, and determine the calculated mean value as the first correlation index.

In order to ensure the accuracy of the data, thecontrol server 10 may also first determine whether the number of preset persons appearing in the target area within a preset time period is greater than a preset threshold. The preset threshold may be a relationship predetermined according to practical experience. If the number of the preset persons appearing in the target area within the preset time period is greater than the preset threshold value, thecontrol server 10 calculates the average value of all the determined second correlation indexes, and determines the calculated average value as the first correlation index. If the number of the preset persons appearing in the target area within the preset time period is smaller than the preset threshold, thecontrol server 10 calculates the average value of all the determined second correlation indexes, and determines the product of the calculated average value and a preset weight (for example, the ratio of the number of the preset persons appearing in the target area within the preset time period to the preset threshold, or other numerical values larger than 0 and smaller than 1) as the first correlation index.

It is easily understood that if the number of preset persons appearing in the target area within the preset time period is higher than the preset threshold, it indicates that the number of "preset persons appearing in the target area within the preset time period" determined by thecontrol server 10 is large. In this case, the probability of occurrence of abnormal clustering is positively correlated with the average value calculated by thecontrol server 10, and the early warning of abnormal clustering can be performed more accurately, so that thecontrol server 10 can determine the calculated average value as the first correlation index. If the number of the preset persons appearing in the target area within the preset time period is lower than the preset threshold, it indicates that the number of the "preset persons appearing in the target area within the preset time period" determined by thecontrol server 10 is smaller. In this case, the probability of occurrence of abnormal aggregation is relatively small, and the finally obtained abnormal aggregation index should be small. If the calculated mean is determined to be the first correlation index, the subsequently obtained abnormal aggregation index may be high. For example: if the preset personnel appearing in the target area within the preset time period include thepreset personnel 1, the preset personnel 2 and the preset personnel 3, and thepreset personnel 1, the preset personnel 2 and the preset personnel 3 belong to the same group, the numerical value of the second association index between every two preset personnel is higher. If the average of these second correlation indexes is taken as the first correlation index, the value of the first correlation index is also higher, and thus the final abnormality aggregation index is also higher. In order to reduce false alarms caused by a small number of "preset persons present in the target area within a preset time period", thecontrol server 10 may determine a product of the calculated average value and a preset weight (e.g., a ratio of the number of preset persons present in the target area within the preset time period to a preset threshold value, or other numerical value greater than 0 and less than 1) as the first correlation index.

The mean calculated by thecontrol server 10 may be an arithmetic mean or a weighted mean, which is not limited in the embodiment of the present application.

The method for thecontrol server 10 to calculate the area index of the target area is as follows: if the target area is determined to be a preset area (if the target area is already marked as the preset area in the system, the preset area may be an area with a high focus, for example, an area with a high frequency of occurrence of an abnormal clustering phenomenon), thecontrol server 10 determines that the index of the high focus area of the target area is 1. On the contrary, thecontrol server 10 calculates the area index of the target area according to the number of times that the preset person appears in the target area.

It is easily understood that, if the number of times the preset person appears in the target area is higher than the preset number of times, it indicates that the preset person often appears in the target area, and thus, the target area is highly likely to be the preset area. Therefore, thecontrol server 10 may determine that the index of the important region of the target region is 1. If the frequency of the preset personnel appearing in the target area is higher than the preset frequency and lower than the preset frequency, the probability of the preset personnel appearing in the target area is low. Thecontrol server 10 may determine a ratio of the number of times that a preset person appears in the target area to a preset number of times as the area index of the target area.

Optionally, thecontrol server 10 may obtain perception data of preset people obtained in a first preset time period (the first preset time period is a historical time period, and the time duration of the first preset time period may be preset), and divide the obtained perception data into a plurality of groups according to the collection location (that is, one group of perception data corresponds to one region). Subsequently, thecontrol server 10 performs the following operations on each set of sensing data: and determining the number of sensing data included in the group, and determining the determined number as the number of times that the preset person appears in the area corresponding to the group of sensing data. I.e. thecontrol server 10 determines the number of times a preset person is present in an area.

Optionally, thecontrol server 10 may further divide the acquired sensing data into a plurality of groups according to the collection location and the type of the group to which the person belongs. Thus, a set of sensory data corresponds to a group and a region. Subsequently, thecontrol server 10 performs the following operations on each set of sensing data: determining the number of persons in the group who include the perception data, and determining the determined number as the number of times the group of persons appears in the area corresponding to the group of perception data. I.e. thecontrol server 10 determines the number of times members of the same group have appeared in an area. Compared with thecontrol server 10 determining the number of times that the preset person appears in one area, thecontrol server 10 determining the number of times that members of the same group appear in one area can further improve the accuracy of thecontrol server 10 determining the area index of the target area.

Specifically, after the first correlation index and the area index of the target area are determined, thecontrol server 10 calculates the abnormal aggregation index of the target area according to the first correlation index and the area index of the target area.

Illustratively, the anomaly cluster index for the target region satisfies the following equation:

AAI＝α*KP+β*KA

in this formula, AAI denotes an abnormal clustering index of the target region, KP denotes a first correlation index, KA denotes a region index of the target region, α and β are both preset weighting coefficients, and α + β is 100.

And S404, when the abnormal aggregation index of the target area meets the preset condition, thecontrol server 10 determines that the abnormal aggregation phenomenon occurs in the target area.

When determining the abnormal aggregation index of the target area, thecontrol server 10 not only considers the association relationship between all preset persons appearing in the target area within a preset time period, but also considers the possibility that the target area is the preset area, and therefore, the accuracy of the abnormal aggregation index of the target area determined by thecontrol server 10 is high.

Optionally, when the abnormal aggregation index of the target area is greater than a preset value, or when the abnormal aggregation index of the target area shows a gradually increasing trend, thecontrol server 10 determines that the abnormal aggregation phenomenon occurs in the target area.

In summary, thecontrol server 10 can predict the target area in advance, determine the probability of the target area having the abnormal aggregation phenomenon (i.e. the abnormal aggregation index of the target area), and determine whether the target area has the abnormal aggregation phenomenon according to the probability, thereby effectively identifying the abnormal aggregation behavior.

Further optionally, after it is determined that the target area is abnormally aggregated, thecontrol server 10 may further send an alarm message to facilitate attention of related personnel (e.g., security personnel) to effectively take measures to avoid occurrence of a potential dangerous event.

With reference to fig. 4, as shown in fig. 7, after S404, the method for identifying an abnormal cluster according to the embodiment of the present application may further include:

s700, thecontrol server 10 sends out an alarm message.

Optionally, thecontrol server 10 may send an alarm sound to remind the video monitoring worker of paying attention, or send an alarm message to a server of the security system to remind the security worker of paying attention, so that the security worker can effectively take measures to avoid the occurrence of a potential dangerous event. In addition, thecontrol server 10 may also send the abnormal alarm to the staff performing the on-site patrol around the target area through other communication means (for example, short message method, etc.), so that the staff can effectively take measures to avoid the occurrence of the potential dangerous event.

Further optionally, with reference to fig. 4, as shown in fig. 7, the method for identifying an abnormal cluster provided in the embodiment of the present application may further include:

s701, thecontrol server 10 periodically updates at least one of the following items: the method comprises the steps of obtaining a region time sequence relation, the association times between a first preset person and a second preset person, and the times of the preset persons appearing in a target region.

Optionally, thecontrol server 10 may update at least one of the area timing relationship, the association frequency between the first preset person and the second preset person, or the frequency of the preset person appearing in the target area according to the sensing data obtained on the same day (or on two adjacent days, which may be determined according to actual configuration), so that the accuracy of the area timing relationship, the association frequency between the first preset person and the second preset person, and the frequency of the preset person appearing in the target area is effectively ensured, and further, the accuracy of predicting the target area by thecontrol server 10 is improved, and the effectiveness of identifying abnormal aggregation is improved.

The method for identifying abnormal aggregation provided by the embodiment of the present application is described below with reference to specific examples.

Theperson 1 is an illegal group member, and the sensing device located in the area a acquires the sensing data of theperson 1 and transmits the sensing data of theperson 1 to thecontrol server 10 through the background server. Thecontrol server 10 determines the vehicle and the area a currently used by theperson 1 based on the perception data of theperson 1. Then, thecontrol server 10 determines, according to the area time-series relationship (refer to table 1), that the target area that theperson 1 is about to reach within the preset time period is the area b. Next, thecontrol server 10 calculates an abnormal aggregation index of the area b (the specific calculation method may refer to the description of S403 described above). If it is determined that the abnormal aggregation index of the area b satisfies the condition, thecontrol server 10 determines that the illegal group members are abnormally aggregated in the area b for a preset time period. Optionally, after the abnormal aggregation phenomenon occurs in the area b within the preset time length, thecontrol server 10 sends out alarm information, so that relevant personnel can take effective measures to avoid the occurrence of a potential dangerous event.

The scheme provided by the embodiment of the application is mainly introduced from the perspective of a method. To implement the above functions, it includes hardware structures and/or software modules for performing the respective functions. Those of skill in the art would readily appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as hardware or combinations of hardware and computer software. Whether a function is performed as hardware or computer software drives hardware depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.

The embodiment of the application also provides an abnormal aggregation identification device. The abnormality aggregation recognition device may be thecontrol server 10, may be a CPU in thecontrol server 10, may be a control module for behavior prediction in thecontrol server 10, or may be a client for behavior prediction in thecontrol server 10.

Fig. 8 is a schematic structural diagram of an abnormalaggregation recognition apparatus 80 according to an embodiment of the present application. The abnormal cluster recognition means 80 is used to execute the abnormal cluster recognition method shown in fig. 4 or fig. 7. The abnormalityaggregation identification device 80 may include adetermination unit 801 and aprediction unit 802.

A determiningunit 801, configured to determine a vehicle currently used by a preset person and an area where the preset person is currently located. For example, in conjunction with fig. 4, thedetermination unit 801 may be configured to perform S401. The predictingunit 802 is configured to predict a target area where a preset person is about to arrive within a preset time according to the transportation currently used by the preset person and the area where the preset person is currently located, which are determined by the determiningunit 801. For example, in conjunction with fig. 4,prediction unit 802 may be used to perform S402. The determiningunit 801 is further configured to determine an abnormal aggregation index of the target region predicted by the predictingunit 802, where the abnormal aggregation index is used to represent a probability of an abnormal aggregation phenomenon occurring in the target region; and the abnormal aggregation index is used for determining that the target area has the abnormal aggregation phenomenon when meeting the preset condition. For example, in connection with fig. 4, thedetermination unit 801 may be configured to perform S403 and S404.

Optionally, the determiningunit 801 is further configured to determine a regional timing relationship; wherein, the regional time sequence relation is used for representing: starting from a first area, using a vehicle, and reaching a second area within a preset time length; the first area at least comprises an area where preset personnel are located currently. The predictingunit 802 is specifically configured to predict a target area where a preset person is about to arrive within a preset time length according to the area time sequence relationship determined by the determiningunit 801, a vehicle currently used by the preset person, and an area where the preset person is currently located.

Optionally, as shown in fig. 8, the anomalyaggregation identifying apparatus 80 further includes astorage unit 803, and thestorage unit 803 may be used to store the region timing relationship.

Optionally, as shown in fig. 8, the anomalyaggregation identifying device 80 further includes an obtainingunit 804. The obtainingunit 804 is configured to obtain a historical movement track of a person, where the historical movement track includes a plurality of track points, and the historical movement track includes a plurality of track points arranged according to a time sequence, and a time corresponding to each track point in the plurality of track points. Correspondingly, the determiningunit 801 is further configured to determine a vehicle corresponding to each of the multiple trace points acquired by the acquiringunit 804, and determine an area time sequence relationship between two trace points having a time difference within the preset time length and being always the same as the vehicle according to the multiple trace points acquired by the acquiringunit 804, the vehicle corresponding to each of the multiple trace points, and time, where the area time sequence relationship between the two trace points is used for representing: follow one of two track points starts, adopts the vehicle can arrive in predetermineeing for a long time another track point in two track points.

Optionally, the determiningunit 801 is specifically configured to: calculating a first association index, wherein the first association index is used for representing the association relation among all preset persons appearing in the target area within a preset time length; calculating a region index of the target region, wherein the region index is used for representing the possibility that the target region is a preset region; and determining an abnormal aggregation index of the target area according to the first correlation index and the area index.

Optionally, the determiningunit 801 is specifically configured to: determining a second correlation index between every two preset persons appearing in the target area within a preset time length; the second association index is used for representing the association relationship between every two preset persons; and calculating the first correlation index according to all the determined second correlation indexes.

Optionally, the determiningunit 801 is specifically configured to: if the first preset person and the second preset person are determined to be members of the same group, determining that a second association index between the first preset person and the second preset person is 1; the first preset personnel and the second preset personnel are any two preset personnel appearing in the target area within a preset time length; otherwise, determining the association times between the first preset personnel and the second preset personnel, and calculating a second association index between the first preset personnel and the second preset personnel according to the association times; the association frequency is the sum of the co-occurrence frequencies of the areas, and one co-occurrence of one area is that the first preset person and the second preset person appear in one area within the preset time difference.

Optionally, the determiningunit 801 is specifically configured to: if the target area is determined to be the preset area, determining that the area index of the target area is 1; otherwise, calculating the area index of the target area according to the times of the preset personnel appearing in the target area.

Of course, the anomalyaggregation identification device 80 provided by the embodiment of the present application includes, but is not limited to, the above modules.

In actual implementation, the determiningunit 801 and the predictingunit 802 may be implemented by theprocessor 31 shown in fig. 3 calling the program code in thememory 32. The specific implementation process may refer to the description of the abnormal aggregation identification method portion shown in fig. 4 or fig. 7, and is not described herein again.

Another embodiment of the present application further provides a computer-readable storage medium, in which computer instructions are stored, and when the computer instructions are executed on the apparatus for identifying an abnormal aggregation, the apparatus for identifying an abnormal aggregation performs each step performed by the apparatus for identifying an abnormal aggregation in the method flow shown in the foregoing method embodiment.

Another embodiment of the present application further provides a chip system, and the chip system is applied to the abnormal aggregation identification apparatus. The system-on-chip includes one or more interface circuits, and one or more processors. The interface circuit and the processor are interconnected by a line. The interface circuit is configured to receive a signal from a memory of the anomaly aggregation identification device and send the signal to the processor, the signal including computer instructions stored in the memory. When the processor executes the computer instructions, the anomaly cluster identification means performs the steps performed by the anomaly cluster identification means in the method flow illustrated in the above-described method embodiments.

In another embodiment of the present application, a computer program product is also provided, where the computer program product includes instructions that, when executed on an anomaly aggregation identifying apparatus, cause the anomaly aggregation identifying apparatus to perform the steps performed by the anomaly aggregation identifying apparatus in the method flow shown in the above-mentioned method embodiment.

The computer instructions may be stored in or transmitted from one computer-readable storage medium to another computer-readable storage medium, e.g., from one website site, computer, server, or data center, by wired (e.g., coaxial cable, fiber optic, digital subscriber line (DS L)) or wireless (e.g., infrared, wireless, microwave, etc.) means for transmitting computer instructions from one website site, computer, server, or data center to another website site, computer, server, or data center, by way of wire (e.g., optical fiber, digital subscriber line (DS L)) or by way of wireless (e.g., infrared, wireless, microwave, etc.), the computer-readable storage medium may be any solid-state accessible medium, such as a floppy disk, a solid-state accessible medium (e.g., a DVD, a floppy disk, or the like), or any other data storage medium, such as a magnetic disk, a floppy disk, or a magnetic disk, or optical disk, or the like.

The foregoing is only illustrative of the present application. Those skilled in the art can conceive of changes or substitutions based on the specific embodiments provided in the present application, and all such changes or substitutions are intended to be included within the scope of the present application.

Claims

1. An abnormal aggregation identification method, comprising:

determining a vehicle currently used by a preset person and a current area of the preset person;

predicting a target area to be reached by the preset personnel within a preset time according to a vehicle currently used by the preset personnel and the area where the preset personnel are currently located;

determining an abnormal aggregation index of the target area, wherein the abnormal aggregation index is used for representing the probability of the abnormal aggregation phenomenon of the target area;

and when the abnormal aggregation index meets a preset condition, determining that the target area has an abnormal aggregation phenomenon.

2. The abnormal aggregation identification method according to claim 1, wherein the predicting, according to a vehicle currently used by the preset person and a current area where the preset person is located, a target area to be reached by the preset person within a preset time period comprises:

determining a regional time sequence relation; wherein the regional timing relationship is used to characterize: starting from a first area, using a vehicle, and reaching a second area within the preset time length; the first area at least comprises an area where the preset personnel are located currently;

and predicting a target area to be reached by the preset personnel within a preset time according to the area time sequence relation, the vehicles currently used by the preset personnel and the current area of the preset personnel.

3. The method according to claim 1 or 2, wherein the determining the abnormal clustering index of the target area comprises:

calculating a first association index, wherein the first association index is used for representing the association relation among all the preset persons appearing in the target area within the preset time length;

calculating a region index of the target region, wherein the region index is used for representing the possibility that the target region is a preset region;

and determining the abnormal aggregation index of the target area according to the first correlation index and the area index.

4. The anomaly aggregation identification method according to claim 3, wherein said calculating a first relevance index comprises:

determining a second correlation index between every two preset persons appearing in the target area within the preset time period; the second association index is used for representing the association relationship between every two preset persons;

and calculating the first correlation index according to all the determined second correlation indexes.

5. The abnormal aggregation identification method of claim 3, wherein the calculating the regional index of the target region comprises:

if the target area is determined to be the preset area, determining that the area index of the target area is 1;

otherwise, calculating the area index of the target area according to the times of the preset personnel appearing in the target area.

6. An abnormal aggregation identification apparatus, comprising:

the device comprises a determining unit, a judging unit and a judging unit, wherein the determining unit is used for determining a vehicle currently used by a preset person and a region where the preset person is currently located;

the prediction unit is used for predicting a target area to be reached by the preset personnel within a preset time length according to the transportation means currently used by the preset personnel and the current area of the preset personnel determined by the determination unit;

the determining unit is further configured to determine an abnormal aggregation index of the target region predicted by the predicting unit, where the abnormal aggregation index is used to represent a probability of an abnormal aggregation phenomenon occurring in the target region; and the abnormal aggregation index is used for determining that the target area has an abnormal aggregation phenomenon when meeting a preset condition.

7. The abnormal aggregation identification device according to claim 6,

the determining unit is further configured to determine a regional timing relationship; wherein the regional timing relationship is used to characterize: starting from a first area, using a vehicle, and reaching a second area within the preset time length; the first area at least comprises an area where the preset personnel are located currently;

the prediction unit is specifically configured to predict a target area to be reached by the preset person within a preset time according to the area time sequence relationship determined by the determination unit, the transportation currently used by the preset person, and the area where the preset person is currently located.

8. The apparatus according to claim 6 or 7, wherein the determining unit is specifically configured to:

9. An apparatus for identifying an anomaly cluster, the apparatus comprising a memory and a processor; the memory and the processor are coupled; the memory for storing computer program code, the computer program code comprising computer instructions; the apparatus for identifying an anomaly cluster, when executed by the processor, performs the method for identifying an anomaly cluster as claimed in any one of claims 1 to 5.

10. A computer-readable storage medium comprising computer instructions which, when run on an anomaly cluster identification apparatus, cause the anomaly cluster identification apparatus to perform the anomaly cluster identification method of any one of claims 1-5.