Methods, devices, electronic equipment and storage media for detecting abnormal populations

Abstract

The present application proposes a method, device, electronic device and storage medium for detecting abnormal crowds, including obtaining real-time signaling sequence data of terminal devices at the current moment of N geographic grids; for any geographic grid in the N geographic grids, according to the moment type of the current moment, determining the historical signaling sequence data corresponding to the real-time signaling sequence data of any geographic grid; splicing the real-time signaling sequence data with the historical signaling sequence data to generate sequence data to be detected; based on the sequence data to be detected of each geographic grid, determining the abnormal geographic grid with abnormal crowds from the N geographic grids. In the present application, the difficulty of detecting abnormal crowd gathering is reduced, based on the signaling data, the detection efficiency is improved while realizing high-coverage abnormal crowd gathering detection, which is independent of video surveillance equipment and reduces the detection cost, thereby realizing efficient, timely, cost-effective and wide-coverage detection of abnormal crowd gathering.

Description

Technical Field

[0001] The present application relates to the field of data statistical analysis, and in particular to a method, device, electronic device and storage medium for detecting abnormalities in a crowd. Background Art

[0002] In the current social management process, the abnormal gathering of people is often accompanied by the occurrence of social problems, and the abnormal gathering of people needs to be paid special attention.

[0003] In the related art, the detection of abnormal crowd gathering depends on video surveillance equipment. After the video surveillance equipment collects video image information, technical analysis is performed to obtain the abnormal crowd gathering situation. Since the deployment cost of video surveillance equipment is high and the timeliness of obtaining the abnormal crowd gathering situation by collecting video image information is low, how to achieve efficient, real-time and cost-effective abnormal crowd gathering detection has become a problem that needs to be solved. Summary of the invention

[0004] The present application aims to solve one of the technical problems in the related art at least to some extent.

[0005] To this end, the first aspect of the present application proposes a method for detecting abnormalities in a crowd.

[0006] The second aspect of the present application also proposes a crowd abnormality detection device.

[0007] A third aspect of the present application provides an electronic device.

[0008] A fourth aspect of the present application provides a computer-readable storage medium.

[0009] A fifth aspect of the present application provides a computer program product.

[0010] The first aspect of the present application proposes a method for detecting crowd anomalies, comprising: obtaining real-time signaling sequence data of terminal devices at the current moment of N geographic grids; for any geographic grid among the N geographic grids, determining the historical signaling sequence data corresponding to the real-time signaling sequence data of the any geographic grid according to the moment type of the current moment; splicing the real-time signaling sequence data with the historical signaling sequence data to generate sequence data to be detected; based on the sequence data to be detected of each geographic grid, determining abnormal geographic grids with crowd anomalies from the N geographic grids.

[0011] In addition, the crowd anomaly detection method proposed in the first aspect of the present application may also have the following additional technical features:

[0012] According to an embodiment of the present application, after determining the abnormal geographical grid with abnormal population from the N geographical grids, it also includes: obtaining the mean and variance of the sequence data to be detected of the abnormal geographical grid, and determining the geographical grid abnormality judgment value based on the mean and the variance; obtaining the number of people at the abnormal geographical grid at the current moment, and selecting the abnormal geographical grid whose number of people at the current moment is greater than the abnormal judgment value as the final abnormal geographical grid.

[0013] According to an embodiment of the present application, the crowd anomaly detection method further includes: acquiring coordinate information of the abnormal geographic grid; and performing geographic grid aggregation processing on the abnormal geographic grid based on the coordinate information to generate an abnormal crowd area.

[0014] According to one embodiment of the present application, the abnormal geographic grids are subjected to geographic grid aggregation processing based on the coordinate information to generate an abnormal population area, including: determining the center point coordinates of the abnormal geographic grids based on the coordinate information of the abnormal geographic grids; determining the distance between any two abnormal geographic grids based on the center point coordinates; grouping the abnormal geographic grids based on the distance and the coordinate information of the abnormal geographic grids to form an aggregated geographic grid group; and performing geographic grid aggregation processing on the abnormal geographic grids belonging to the same group to generate the abnormal population area.

[0015] According to one embodiment of the present application, the abnormal geographic grids are grouped according to the distance and the coordinate information of the abnormal geographic grids to form an aggregated geographic grid group, including: selecting an abnormal geographic grid pair whose distance is less than a preset distance threshold, wherein any two abnormal geographic grids constitute an abnormal geographic grid pair; based on the coordinate information of the abnormal geographic grids, determining one or more aggregated geographic grid groups that are continuous in spatial position from the abnormal geographic grid pair, wherein the aggregated geographic grid group includes one or more abnormal geographic grid pairs.

[0016] According to one embodiment of the present application, before obtaining the real-time signaling sequence data of the terminal device at the current moment of N geographic grids, it also includes: obtaining the historical signaling data of the N geographic grids; for any geographic grid among the N geographic grids, dividing the historical signaling data of the any geographic grid according to a preset time interval to generate multiple signaling sub-sequence data; identifying the time type of the moment of the signaling sub-sequence data, and splicing the signaling sub-sequence data belonging to the same time category in chronological order from early to late to obtain the historical signaling sequence data of each time type.

[0017] According to one embodiment of the present application, the crowd anomaly detection method also includes: obtaining the collection time length corresponding to the historical signaling data within any geographic grid; in response to the collection time length being greater than a preset time length, performing time type identification based on a clustering algorithm; in response to the collection time length being less than or equal to the preset time length, performing time type identification according to classification rules.

[0018] According to an embodiment of the present application, after generating the abnormal crowd area, the method further includes: acquiring the maximum coordinate and the minimum coordinate in the abnormal crowd area as the area boundary of the abnormal crowd area.

[0019] According to one embodiment of the present application, after generating the abnormal crowd area, it also includes: obtaining geographical grids belonging to the abnormal crowd area, and counting the number of people in each geographical grid belonging to the abnormal crowd area at the current moment, as well as the historical number of people for a set period of time; splicing the number of people at the current moment and the historical number of people to generate number sequence data of the abnormal crowd area.

[0020] The second aspect of the present application also proposes a crowd anomaly detection device, including: an acquisition module, used to acquire real-time signaling sequence data of terminal devices at the current moment of N geographical grids; the acquisition module is also used to determine, for any geographical grid among the N geographical grids, according to the moment type of the current moment, the historical signaling sequence data corresponding to the real-time signaling sequence data of any geographical grid; a splicing module, used to splice the real-time signaling sequence data with the historical signaling sequence data to generate sequence data to be detected; a detection module, used to determine abnormal geographical grids with crowd anomalies from the N geographical grids based on the sequence data to be detected of each geographical grid.

[0021] The crowd anomaly detection device proposed in the second aspect of the present application may also have the following additional technical features:

[0022] According to one embodiment of the present application, the crowd anomaly detection device also includes: a calculation module, used to obtain the mean and variance of the sequence data to be detected of the abnormal geographical grid, and determine the geographical grid anomaly judgment value based on the mean and the variance; an abnormal geographical grid determination module, used to obtain the number of people in the abnormal geographical grid at the current moment, and select the abnormal geographical grid whose number of people at the current moment is greater than the anomaly judgment value as the final abnormal geographical grid.

[0023] According to one embodiment of the present application, the crowd anomaly detection device further includes: a coordinate acquisition module, used to obtain coordinate information of the abnormal geographic grid; an aggregation module, used to perform geographic grid aggregation processing on the abnormal geographic grid based on the coordinate information to generate an abnormal crowd area.

[0024] According to one embodiment of the present application, the aggregation module is also used to: determine the center point coordinates of the abnormal geographic grid based on the coordinate information of the abnormal geographic grid; determine the distance between any two abnormal geographic grids based on the center point coordinates; group the abnormal geographic grids according to the distance and the coordinate information of the abnormal geographic grids to form an aggregated geographic grid group; perform geographic grid aggregation processing on the abnormal geographic grids belonging to the same group to generate the abnormal population area.

[0025] According to one embodiment of the present application, the aggregation module is also used to: select the abnormal geographic grid pair whose distance is less than a preset distance threshold, wherein any two abnormal geographic grids constitute an abnormal geographic grid pair; based on the coordinate information of the abnormal geographic grids, determine one or more continuous aggregated geographic grid groups in spatial position from the abnormal geographic grid pair, wherein the aggregated geographic grid group includes one or more abnormal geographic grid pairs.

[0026] According to one embodiment of the present application, the crowd anomaly detection device also includes: the acquisition module, which is also used to acquire the historical signaling data of the N geographic grids; the segmentation module, which is used to segment the historical signaling data of any one of the N geographic grids according to a preset time interval to generate multiple signaling sub-sequence data; the splicing module is also used to identify the time type of the moment of the signaling sub-sequence data, and splice the signaling sub-sequence data belonging to the same time category in chronological order from early to late to obtain the historical signaling sequence data of each time type.

[0027] According to one embodiment of the present application, the crowd anomaly detection device also includes: the acquisition module, which is also used to obtain the collection time length corresponding to the historical signaling data within any geographic grid; the identification module, which is used to perform time type identification based on a clustering algorithm in response to the collection time length being greater than a preset time length; the identification module is also used to perform time type identification according to classification rules in response to the collection time length being less than or equal to the preset time length.

[0028] According to an embodiment of the present application, the crowd anomaly detection device further includes: a boundary generation module, which is used to obtain the maximum coordinates and the minimum coordinates in the crowd anomaly area as the area boundary of the crowd anomaly area.

[0029] According to one embodiment of the present application, the crowd anomaly detection device also includes: a number acquisition module, which is used to obtain the geographical grids belonging to the crowd anomaly area, and count the number of people in each geographical grid belonging to the crowd anomaly area at the current moment, as well as the historical number of people for a set period of time; the splicing module is also used to splice the number of people at the current moment and the historical number of people to generate the number sequence data of the crowd anomaly area.

[0030] The third aspect of the present application proposes an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the crowd anomaly detection method proposed in the first aspect above.

[0031] The fourth aspect of the present application proposes a computer-readable storage medium, wherein the computer instructions are used to enable the computer to execute the crowd anomaly detection method proposed in the first aspect.

[0032] The fifth aspect of the present application proposes a computer program product, including a computer program, which, when executed by a processor, implements the crowd anomaly detection method proposed in the first aspect.

[0033] The crowd abnormality detection method and device proposed in the present application obtains the real-time signaling sequence data of the terminal equipment at the current moment of N geographic grids, and determines the historical signaling sequence data corresponding to each geographic grid in the N geographic grids according to the time type, splices the real-time signaling sequence data with the historical signaling sequence data to generate the sequence data to be detected, and obtains the abnormal geographic grids in the N geographic grids where the crowd abnormally gathers based on the sequence data to be detected. In the present application, the geographic location information is divided into N geographic grids, which reduces the difficulty of abnormal crowd gathering detection. Through the signaling data, it is possible to judge whether the crowd has abnormal gathering, improve the detection efficiency, and realize high-coverage abnormal crowd gathering detection, which is not dependent on video surveillance equipment, reduces the cost of detection, and thus realizes efficient, timely, cost-effective and wide-coverage abnormal crowd gathering detection.

[0034] It should be understood that the contents described in this application are not intended to identify the key or important features of the embodiments of this application, nor are they intended to limit the scope of this application. Other features of this application will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0036] Figure 1 A flow chart of a method for detecting abnormal crowds according to an embodiment of the present application;

[0037] Figure 2 A flow chart of a method for detecting abnormal crowds according to another embodiment of the present application;

[0038] Figure 3 A flow chart of a method for detecting abnormal crowds according to another embodiment of the present application;

[0039] Figure 4 A flow chart of a method for detecting abnormal crowds according to another embodiment of the present application;

[0040] Figure 5 A flow chart of a method for detecting abnormal crowds according to another embodiment of the present application;

[0041] Figure 6 A flow chart of a method for detecting abnormal crowds according to another embodiment of the present application;

[0042] Figure 7 A flow chart of a method for detecting abnormal crowds according to another embodiment of the present application;

[0043] Figure 8 This is a schematic diagram of the structure of a crowd anomaly detection device according to an embodiment of the present application;

[0044] Fig. 9 This is a structural schematic diagram of a crowd anomaly detection device according to another embodiment of the present application;

[0045] Fig.10 It is a structural schematic diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0046] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0047] The following describes the crowd anomaly detection method, device, electronic device and storage medium of the embodiments of the present application with reference to the accompanying drawings.

[0048] Figure 1 FIG. 1 is a flow chart of a method for detecting abnormal crowds according to an embodiment of the present invention. Figure 1 As shown, the method includes:

[0049] S101, obtaining real-time signaling sequence data of terminal devices in N geographic grids at the current moment.

[0050] At present, there are terminal devices with a very high penetration rate, such as mobile devices, wearable devices and other portable electronic devices. During use, the terminal device can send signaling data to the corresponding receiving device in real time or at a set interval, so as to realize the interaction of information or data between the terminal device and the receiving device through signaling data. The signaling data may include device information, time information, spatial information, number of signaling data users, etc. of the terminal device. Among them, the spatial information is a character string that can represent the geographic location and is generated by encoding the algorithm, such as a one-dimensional character string generated by encoding the geographic location through an address encoding algorithm (geohash), which can correspond to a geographic grid of a set area on the map. The number of signaling data users can be understood as the number of users obtained based on the number of terminal devices with which signaling data interaction exists.

[0051] Furthermore, based on the high penetration rate of terminal devices, the signaling data corresponding to the terminal devices can be used as alternative information of the users, such as the user's movement trajectory, the number of users in a certain area or time range, the time when the terminal devices send or receive signaling data, etc.

[0052] It should be noted that the time interval for counting the number of people using signaling data can be set. For example, if the time interval is set to 5 minutes, the signaling data will be counted every 5 minutes.

[0053] In general, for a certain geographical location, the longitude and latitude coordinates can usually be used to mark the geographical location, wherein the longitude and latitude coordinates are a two-dimensional array, which does not represent a certain point on the surface of the earth, but refers to a general area, and the range of the represented area is directly related to the accuracy of the longitude and latitude values. Therefore, the area can be divided into multiple geographic grids based on the algorithm and the longitude and latitude of a certain area, such as the geohash algorithm. The geohash algorithm can divide a map or an area on a map, and then divide the map or an area on a map into rectangular blocks by using base32 encoding, wherein the rectangular blocks obtained by division are the geographic grids in the embodiment of the present application.

[0054] In the embodiment of the present application, the signaling data of the current moment of N geographic grids can be determined as the real-time signaling data of N geographic grids. By acquiring and analyzing the signaling data, the relevant information of the user of the terminal device can be indirectly acquired. For example, based on the spatial information and time information contained in the signaling data, the geographical location of the user at a certain moment can be acquired.

[0055] In order to achieve real-time detection of abnormal crowd gathering, the signaling data of the terminal device can be obtained through the service provider of the terminal device, and then the effective information that can be used to determine whether the crowd is abnormally gathered can be obtained, such as geographic location information, time information, and the amount of signaling data, etc. Furthermore, by analyzing the effective information in the signaling data, the number of people in a certain geographic location at a certain time can be obtained, and combined with the historical crowd gathering data of the geographic location, it can be determined whether there is abnormal gathering.

[0056] In order to better count the geographic location information of the terminal device, multiple encoding methods can be used to encode the geographic location information, and then obtain characters that can represent the geographic location. For example, the geohash algorithm can be used to encode the latitude and longitude data of a geographic location into a one-dimensional string.

[0057] In an embodiment of the present application, the longitude and latitude of the earth's surface are encoded using the geohash algorithm, and the earth's surface can be divided into N geographic grids. Based on the set accuracy of the geohash algorithm, a fixed area size is set for each geographic grid, and each geographic grid corresponds to a geographic location on the earth's surface.

[0058] In order to detect abnormal crowd gathering in N geographic grids at the current moment, it is necessary to obtain information on crowd gathering in N geographic grids at the current moment, including geographic grid information, time information and the amount of signaling data, wherein the amount of signaling data can be understood as the number of terminal devices that have sent signaling data, and the time information and the amount of signaling data correspond to the geographic grid information.

[0059] Furthermore, the real-time signaling data can be preprocessed to generate real-time signaling sequence data. Optionally, according to the set data format, the acquired geographic grid information, the time information corresponding to the geographic grid information, and the amount of signaling data are converted into a format, thereby acquiring the real-time signaling sequence data of the terminal device at the current moment.

[0060] For example, if we need to obtain real-time signaling sequence data of 50 geographic grids and collect signaling data statistics of 50 geographic grids every 5 minutes, we can obtain the number of people gathering in each geographic grid at the current moment as P. wt , where w is the geographic grid identity document (id) and t is the time accurate to the minute. Furthermore, the real-time signaling sequence data can be understood as a string of P wt sequence.

[0061] S102 : for any geographical grid among the N geographical grids, according to the moment type of the current moment, determine the historical signaling sequence data corresponding to the real-time signaling sequence data of any geographical grid.

[0062] The gathering of crowds is affected by many factors. For example, different geographical locations have different crowd distributions. For another example, the same geographical location has different crowd distributions at different times. In order to achieve more accurate detection of abnormal crowd gatherings, it is necessary to compare the current crowd gathering situation with the historical crowd gathering situation.

[0063] Generally speaking, in the same geographical location and at the same time, the gathering of people is roughly similar. For example, for a large shopping mall, a longitudinal comparison of the crowd gathering of the mall in the past three years shows that the mall has a large flow of people on non-working days of the week, and through statistical static distribution, it can be seen that the historical gathering of people in the large shopping mall has a range of values. From this, it can be judged that in the future, if there is no abnormal gathering of people, the flow of people in the mall on weekends will remain within the range value obtained based on historical data.

[0064] In the embodiment of the present application, for each of the N geographic grids, there may be historical signaling sequence data of different time types. The real-time signaling sequence data of the terminal device at the current moment of the N geographic grids includes time information, which can be identified according to the time type set by the historical signaling sequence data, for example, the Spring Festival holiday can be classified as a holiday, Monday to Friday can be classified as a working day, and so on.

[0065] Furthermore, according to the time type corresponding to the time information in the real-time signaling sequence data of the terminal devices of the N geographic grids at the current moment, the historical signaling sequence data of the same time type corresponding to the time type to which the real-time signaling sequence data of the N geographic grids at the current moment belongs can be obtained from the historical signaling sequence data corresponding to each time type in the N geographic grids.

[0066] S103, splicing the real-time signaling sequence data with the historical signaling sequence data to generate sequence data to be detected.

[0067] In order to more accurately identify whether the crowd gathering situation carried in the real-time signaling sequence data at the current moment is abnormal, it is necessary to compare the real-time signaling sequence data with the historical signaling sequence data.

[0068] In order to reduce the difficulty of identifying real-time signaling sequence data, the real-time signaling sequence data can be spliced ​​with the historical signaling sequence data, so that the crowd gathering information carried in the historical signaling sequence data and the real-time signaling sequence data can be obtained simultaneously during the detection process.

[0069] Among them, the historical crowd gathering conditions carried in the historical signaling sequence data are used as the detection standard, and then new sequence data containing historical signaling sequence data and real-time signaling sequence data are generated.

[0070] Furthermore, the real-time signaling sequence data can be spliced ​​after the historical signaling sequence data, so that during the detection process, the historical data of normal crowd gathering in the historical signaling sequence data can be obtained first, and then after reading the real-time signaling sequence data in the sequence data, the crowd gathering situation carried in the real-time signaling sequence data can be directly compared with the historical crowd gathering situation in the historical signaling sequence data obtained in advance, effectively shortening the comparison time.

[0071] Furthermore, the historical signaling sequence data is spliced ​​with the real-time signaling sequence data to generate the sequence data to be detected.

[0072] For example, the number of people gathered in 50 geographic grids obtained every 5 minutes is P wt , then the real-time signaling sequence data at the current moment is a string of P wt If the time type of the current moment is set to q, the historical signaling sequence data can be obtained as P′ w q, further, the real-time signaling sequence data P wt Splice to historical signaling sequence data P′ w After q, the sequence data to be detected is P w .

[0073] S104, based on the sequence data to be detected of each geographic grid, determining an abnormal geographic grid with abnormal population from the N geographic grids.

[0074] For the sequence data to be detected of a certain geographic grid among N geographic grids, the real-time signaling sequence data contained therein is compared with the historical signaling sequence data to determine whether there is abnormal crowd gathering in the geographic grid at the current moment.

[0075] For example, for the detection of abnormal crowd gathering in geographic grid A during summer holidays, in the historical signaling sequence data of geographic grid A, the corresponding large shopping mall has a flow of 5,000 to 10,000 people per day during summer holidays. The real-time signaling sequence data of the terminal device of geographic grid A at the current moment is obtained, where the time type at the current moment is a summer holiday, and the flow of people on that day reaches 20,000 people. Based on this situation, it can be judged that abnormal crowd gathering occurs in geographic grid A at the current moment.

[0076] Optionally, a model for detecting abnormal crowd gatherings may be set, and the sequence data to be detected may be input into the model for detecting abnormal crowd gatherings. Based on the output result of the model, it may be determined whether abnormal crowd gatherings occur in the geographic grid corresponding to the sequence data to be detected.

[0077] In the embodiment of the present application, both historical signaling sequence data and real-time signaling sequence data belong to time series data. Generally, time series data has seasonality and trend. When performing anomaly detection on time series data, anomaly detection needs to be performed based on the two basic characteristics of time series data. A certain time series data cannot be treated as an isolated sample point. Furthermore, a time series anomaly detection algorithm (Seasonal Hybrid Extreme Studentized Deviate, SH-ESD) can be used.

[0078] For the sequence data to be detected corresponding to a certain geographic grid, by analyzing the historical crowd gathering situation carried by the historical signaling sequence data, the periodicity and trend of the historical crowd gathering of the geographic grid can be obtained. The crowd gathering situation in the real-time signaling sequence data of the geographic grid at the current moment corresponds to the historical signaling sequence data in the same sequence data to be detected in terms of time period.

[0079] Furthermore, the crowd gathering situation carried in the real-time signaling sequence data of the geographic grid is compared with the historical crowd gathering trend of the geographic grid. If the crowd gathering situation of the geographic grid at the current moment is within the range of the historical crowd gathering trend, it can be judged that there is no abnormal crowd gathering in the geographic grid at the current moment. If the crowd gathering situation of the geographic grid at the current moment exceeds the upper limit of the range of the historical crowd gathering trend or is lower than the lower limit of the range of the historical crowd gathering trend, it can be judged that there is abnormal crowd gathering in the geographic grid at the current moment, and the geographic grid is then determined to be an abnormal geographic grid.

[0080] The crowd anomaly detection method proposed in the present application obtains the real-time signaling sequence data of the terminal device at the current moment of N geographic grids, and determines the historical signaling sequence data corresponding to each geographic grid in the N geographic grids according to the time type, splices the real-time signaling sequence data with the historical signaling sequence data to generate the sequence data to be detected, and obtains the abnormal geographic grids in the N geographic grids where the crowd abnormally gathers based on the sequence data to be detected. In the present application, the geographic location information is divided into N geographic grids, which reduces the difficulty of abnormal crowd gathering detection. Through the signaling data, it is possible to judge whether the crowd has abnormal gathering, improve the detection efficiency, and realize high-coverage abnormal crowd gathering detection. It does not rely on video surveillance equipment, reduces the cost of detection, and thus realizes efficient, timely, cost-effective and wide-coverage abnormal crowd gathering detection.

[0081] Based on the above embodiment, the acquisition of historical signaling sequence data can be combined with Figure 2 Further understanding, Figure 2 FIG. 1 is a flow chart of a method for detecting abnormal crowds according to another embodiment of the present application. Figure 2 As shown, the method includes:

[0082] S201, obtaining historical signaling data of N geographical grids.

[0083] In general, crowd gatherings fluctuate periodically over time. For example, for a park corresponding to a certain geographic grid, there are fewer people in the park during the day and more people at night on weekdays, and there are more people in the park during holidays throughout the day. The number of people in the park in summer is much higher than that in winter, and so on. In order to more accurately detect abnormal crowd gatherings at a certain geographic grid at the current moment, it is necessary to obtain the historical signaling data of crowd gatherings in the geographic grid to analyze the cycles and trends of historical crowd gatherings in the geographic grid under normal circumstances.

[0084] The historical signaling data of N geographic grids may be obtained from the service provider of the terminal device, wherein the historical signaling data contains device information of the terminal device, historical time information, historical spatial information, number of people using the historical signaling data, etc. The spatial information may be a character string that can represent the geographic location and is generated by encoding through an algorithm, such as a one-dimensional character string generated by encoding the geographic location through the geohash algorithm proposed in the above embodiment.

[0085] In the embodiment of the present application, signaling data corresponding to N geographic grids within a continuous set time range in the past may be obtained, that is, historical signaling data corresponding to the N geographic grids.

[0086] For example, set to obtain the historical signaling data corresponding to the time span from May 1 to May 31 of 50 geographic grids, where the crowd gathering situation of the 50 geographic grids is counted every 5 minutes, and the number of people gathered corresponding to the 50 geographic grids P′ can be obtained. wt , where w is the geographic grid id and t is the time accurate to the minute. Further, the acquired historical signaling data is sorted in chronological order to obtain a string of P′ wt The sequence is the historical signaling data corresponding to 50 geographic grids with a collection time length from May 1 to May 31.

[0087] S202 : For any geographical grid among the N geographical grids, data segmentation is performed on the historical signaling data of any geographical grid according to a preset time interval to generate a plurality of signaling subsequence data.

[0088] In order to detect abnormal crowd gathering, the acquired historical signaling data needs to be classified and preprocessed. Since the method of crowd anomaly detection in the embodiment of the present application is based on the different crowd gathering situations of time series data statistics, it is necessary to segment the historical signaling data according to preset time intervals, such as data segmentation by day.

[0089] For a certain geographic grid among the N geographic grids, the historical signaling data corresponding to the geographic grid is segmented according to a preset time interval to generate multiple signaling subsequence data. For example, the time range of the historical signaling data corresponding to the geographic grid is set to May 1 to May 31, and the historical signaling data is segmented by day, and 31 signaling subsequence data can be obtained.

[0090] It should be noted that, in the N geographical grids, each geographical grid needs to perform data segmentation on its corresponding historical signaling data at a preset time interval.

[0091] For example, for 50 geographic grids (W1, W2, ..., W 50 ), set to obtain the historical signaling data P' corresponding to each geographic grid from May 1 to May 31 wt , the acquired historical signaling data P′ can be wt The composed sequence is divided into data by day, and 31 signaling subsequence data P′ are obtained wdt , where d represents the date and t does not contain the date.

[0092] For example, for 50 geographic grids (W1, W2, ..., W 50 ), set to obtain the historical signaling data P' corresponding to each geographic grid from May 1 to May 31 wt, the acquired historical signaling data P′ can be wt The composed sequence is segmented at a set time interval of 4 hours, and 186 signaling subsequence data P′ are obtained. wdt ', where d represents the date and t does not contain the date.

[0093] S203, identify the time type of the signaling subsequence data, and splice the signaling subsequence data belonging to the same time category in chronological order from early to late, to obtain historical signaling sequence data of each time type.

[0094] In an embodiment of the present application, for a certain geographical grid among N geographical grids, after obtaining multiple signaling sub-sequence data generated by dividing the historical signaling data of the geographical grid, it is necessary to classify the multiple signaling sub-sequence data based on the time type, and then generate historical signaling sequence data corresponding to the time type in the geographical grid.

[0095] Furthermore, a preset collection time length corresponding to the historical signaling data of the geographic grid can be set, and based on the comparison result between the preset time length and the actual collection length of the historical signaling data of the geographic grid, the obtained multiple signaling subsequence data are classified. It can be understood that the number of historical signaling data collected when the collection time length is greater than the preset collection time length can meet the number of samples required by the clustering adaptive classification method.

[0096] Optionally, in response to the acquisition time length being greater than a preset time length, time type recognition is performed based on a clustering algorithm.

[0097] For a certain geographical grid among N geographical grids, when the actual collection time length of the historical signaling data corresponding to the geographical grid is greater than the preset collection time length, a clustering algorithm can be used to enable the multiple signaling sub-sequence data of the geographical grid to be adaptively classified, thereby realizing the classification of the multiple signaling sub-sequence data of the geographical grid.

[0098] For example, the preset collection time length corresponding to the historical signaling data is set to one year equal to 365 days, and for a certain geographical grid, its historical signaling data P′ is set to wt The collection time is longer than one year, and the historical signaling data P′ wt Data is cut by day, and signaling subsequence data P′ with a quantity greater than 365 is obtained. wdt , then all signaling subsequence data P′ are clustered wdtClassify into k time types, where the number of time types k is automatically generated by the clustering model based on the time types of the identified signaling subsequence data. Further, the signaling subsequence data under the k time types are spliced ​​in chronological order to generate the historical signaling sequence data P′ corresponding to the k time types. w 1,P′ w 2,…,P′ w k.

[0099] It should be noted that the classification of signaling subsequence data through clustering algorithms can make the classification results of signaling subsequence data cover a wider range. It can be understood that the classification of signaling subsequence data through clustering methods can identify more time types other than the usual time types in the clustering process. For example, the flow of people on weekends in severe winter and midsummer is less than that on weekends in warm spring and cool autumn. The clustering algorithm can more easily obtain the rules of crowd gathering of this type, classify severe winter and midsummer into one time type, and classify warm spring and cool autumn into another time type.

[0100] The clustering algorithm may be a density-based clustering algorithm, such as the DBSCAN algorithm (Density-Based Spatial Clustering of Applications with Noise), or a partition-based clustering algorithm, such as the K-means clustering algorithm (K-MEANS algorithm), and the like.

[0101] Clustering multiple signaling subsequence data, identifying the time type in the clustering process, and then classifying the signaling subsequence data into each acquired time type, and generating a mapping relationship from time to time type.

[0102] Optionally, in response to the acquisition time length being less than or equal to a preset time length, time type identification is performed according to a classification rule.

[0103] For a certain geographic grid among N geographic grids, when the collection time length of the historical signaling data corresponding to the geographic grid is less than or equal to the preset time length, the number of historical signaling data collected in this scenario is relatively small, which does not meet the number of samples required by the adaptive classification method of clustering. In this scenario, it is necessary to manually classify the signaling subsequence data. A classification rule can be set, which includes multiple manually set time types and relevant features of the signaling subsequence data that can be classified into a certain time type. Based on the classification rule, the multiple signaling subsequence data corresponding to the geographic grid are classified in terms of time type, and then a mapping relationship from time to time type is generated.

[0104] For example, the preset collection time length corresponding to the historical signaling data is set to one year equal to 365 days, and for a certain geographical grid, its historical signaling data P′ is set to wt The collection period is from May 1 to May 31. wt The data is cut by day, and 31 signaling subsequence data P′ are obtained. wdt Since 31 is less than 365, it is necessary to classify the signaling subsequence data based on the set classification rules. Further, set 31 signaling subsequence data P′ wdt The time types to be classified are three time types: working days, holidays, and non-working and / or holidays. In the collection time length from May 1 to May 31, there are 20 working days, 5 holidays, and 6 non-working and / or holidays. According to the set rules, the time type of working days contains 20 signaling subsequence data, the time type of holidays contains 5 signaling subsequence data, and the time type of non-working and / or holidays contains 6 signaling subsequence data. Further, multiple signaling subsequence data under each time type are spliced ​​in chronological order to generate historical signaling sequence data P′ under the working day time type. w 1. Historical signaling sequence data P′ under holiday time type w 2 and historical signaling sequence data P′ under non-working and / or holiday time types w 3.

[0105] It should be noted that the above classification and preprocessing method for historical signaling data of a certain geographic grid is applicable to all N geographic grids. Each geographic grid can obtain its corresponding historical signaling sequence data containing one or more time types based on the method in the above embodiment.

[0106] Furthermore, assuming that each of the N geographic grids outputs historical signaling sequence data of k time types, after performing the above operation on each geographic grid, M sets of historical signaling sequence data can be obtained, and the formula is as follows:

[0107]

[0108] The crowd anomaly detection method proposed in the present application divides the historical signaling data of N geographical grids acquired at preset time intervals to generate signaling subsequence data, and classifies the signaling subsequence data into time types, and further, splices the signaling subsequence data belonging to the same time type in chronological order to obtain the historical signaling sequence data of all time types corresponding to each geographical grid in the N geographical grids. In the present application, the acquisition of historical signaling sequence data is achieved through the classification and preprocessing of historical signaling data, which provides data for obtaining the periodicity and trend of abnormal crowd gathering, thereby improving the accuracy of the detection of abnormal crowd gathering at the current moment, and realizing efficient, timely, cost-effective and wide-coverage detection of abnormal crowd gathering.

[0109] In general, for a certain geographic grid, it is possible to determine whether abnormal crowd gathering occurs in the geographic grid based on the result of comparing the crowd gathering situation at the current moment carried in its real-time signaling sequence data with the historical crowd gathering situation carried in the historical signaling data, and determine all situations that do not conform to the trend and periodicity of crowd gathering obtained by the geographic grid based on the historical signaling data as the detection results of abnormal crowd gathering, such as a sudden increase in the number of people, a sudden decrease in the number of people, etc. Since the detection method for abnormal crowd gathering proposed in the embodiment of the present application is mainly aimed at abnormal gathering situations with a sudden increase in the number of people, after determining the abnormal geographic grid with abnormal crowd gathering in N geographic grids, it is necessary to further perform statistical analysis on the abnormal crowd gathering situation of this part of the abnormal geographic grids to obtain the abnormal geographic grid with a sudden increase in the number of people.

[0110] In order to more accurately locate the range of abnormal crowd gathering, it can be combined with Figure 3 Further understanding, Figure 3 FIG. 1 is a flow chart of a method for detecting abnormal crowds according to another embodiment of the present application. Figure 3 As shown, the method includes:

[0111] S301, training a time series anomaly detection model, and determining abnormal geographic grids with abnormal populations among N geographic grids based on the trained time series anomaly detection model.

[0112] In an embodiment of the present application, for a certain geographical grid, abnormal crowd gathering detection in the geographical grid is achieved by comparing its historical signaling data with the real-time signaling data. It can be understood that abnormal crowd gathering detection is achieved by comparing the crowd gathering situations of the geographical grid at different time points.

[0113] Therefore, a time series anomaly detection model can be set up, and the historical signaling data of N geographic grids can be obtained to train the time series anomaly detection model. For example, the historical signaling data of a certain geographic grid in the past ten years can be used as training samples, where each training sample has a crowd aggregation label, which can indicate whether the training sample is normal. For example, the number 1 can be used to identify a normal training sample, and the number 0 can be used to identify an abnormal training sample.

[0114] Furthermore, the training samples are input into the time series anomaly detection model for training, the predicted crowd aggregation results are compared with the corresponding crowd aggregation in the actual historical signaling data, and the various parameters of the time series anomaly detection model are adjusted based on the comparison results until the final training output results can meet the needs of practical applications.

[0115] Furthermore, the sequence data to be detected of N geographic grids are input into the trained time series anomaly detection model. The model can predict whether each sequence data to be detected has abnormal crowd gathering, and determine the geographic grid corresponding to the sequence data to be detected where abnormal crowd gathering occurs, that is, the abnormal geographic grid confirmed to have abnormal crowd gathering from the N geographic grids.

[0116] S302, obtaining the mean and variance of the sequence data to be detected of the abnormal geographic grid, and determining the geographic grid abnormality judgment value based on the mean and variance.

[0117] In general, the sudden increase of the crowd in a certain place is uncontrollable. Although the periodicity and trend of the sudden increase of the crowd in the geographic grid can be obtained according to the historical number of people gathered in N geographic grids in the embodiment of the present application, there are also normal crowd gatherings in abnormal geographic grids. For example, a large-scale event held for the first time in a large shopping mall, vertically comparing the historical crowd gathering of the mall, the number of people gathered in the mall during this event increased sharply, but horizontally comparing the crowd gathering in similar activities in other shopping malls with similar locations, conditions and other parameters, it can be judged that the crowd gathering of this event in the mall is within the normal gathering range.

[0118] Therefore, it is necessary to further judge the increase in the number of people in abnormal geographic grids and obtain the geographic grids where the abnormal increase occurs. For a certain abnormal geographic grid, its historical signaling data can be statistically analyzed for static distribution probability to obtain the range of normal increase in the population of the geographic grid. Furthermore, the upper limit X of the normal number of people gathered in the history of the geographic grid can be obtained by the following formula:

[0119]

[0120] in, represents the mean of the historical signaling sequence data corresponding to the time type to which the geographic grid belongs at the current moment, and σ represents the variance of the historical signaling sequence data corresponding to the time type to which the geographic grid belongs at the current moment.

[0121] Furthermore, the upper limit X of the number of people in the historical normal gathering of the geographic grid is the geographic grid abnormality judgment value corresponding to the geographic grid. Further, the sequence data to be detected corresponding to all abnormal geographic grids can be obtained, and then the geographic grid abnormality judgment values ​​corresponding to all abnormal geographic grids can be obtained.

[0122] For example, in the above example of 50 geographic grids (W1, W2, …, W 50 ) is used to detect abnormal crowd gatherings, and the geographic grids W1, W7, W 15 ,W 22 ,W 38 Clustering occurs. For the geographic grids W1, W7, W 15 ,W 22 ,W 38 Further screening is performed to obtain the time type q corresponding to the five abnormal geographic grids and the corresponding sequence data to be detected as P w , we can obtain the historical signaling sequence data P′ corresponding to the abnormal geographic grid w q. Further, obtain the historical signaling sequence data P′ w The mean of q and variance σ q , generate the geographic grid anomaly judgment value X corresponding to each abnormal geographic grid q , where the geographic grid anomaly determination value

[0123] S303, obtaining the number of people in the abnormal geographic grid at the current moment, and selecting the abnormal geographic grid whose number of people at the current moment is greater than the abnormal determination value as the final abnormal geographic grid.

[0124] After determining the abnormal geographic grid where the crowd is abnormally gathered, the sequence data to be detected corresponding to each abnormal geographic grid can be obtained. Among them, the sequence data to be detected carries the real-time signaling sequence data of the abnormal geographic grid at the current moment and the historical signaling sequence data of the set time. Further, based on the real-time signaling sequence data at the current moment carried by the sequence data to be detected, the number of people gathered in the abnormal geographic grid at the current moment is obtained.

[0125] The embodiment of the present application is aimed at detecting abnormal crowd gathering in the case of a sudden increase in the crowd. Therefore, the number of people gathered in the acquired abnormal geographical grid at the current moment can be compared with the acquired geographical grid abnormal judgment value, and the abnormal geographical grid whose number of people gathered at the current moment is greater than the corresponding geographical grid abnormal judgment value can be screened out. The increase in the crowd in the abnormal geographical grid can be understood as an abnormal sudden increase.

[0126] Furthermore, the abnormal geographic grid obtained to determine the abnormal crowd gathering situation is determined as the final abnormal geographic grid.

[0127] For example, based on the above example, obtain the abnormal geographic grids W1, W7, W 15 ,W 22 ,W 38 The corresponding sequence data P to be detected w Then, the number of people gathered at the current moment in the five abnormal geographic grids is compared with the corresponding geographic grid abnormality judgment value X q For comparison, we set the five abnormal geographic grids, geographic grids W7, W 15 The current number of people gathered is less than or equal to the corresponding geographic grid anomaly judgment value X q , geographic grid W1,W 22 ,W 38 The number of people gathered at the current moment is greater than the corresponding geographic grid anomaly judgment value X q , we can determine the geographic grid W1,W 22 ,W 38 This is the final abnormal geographic grid for the population surge.

[0128] The crowd anomaly detection method proposed in the embodiment of the present application obtains the abnormal geographic grid where the abnormal crowd gathers through the time series anomaly detection model, and further performs statistical analysis on the obtained abnormal geographic grid, and determines the abnormal geographic grid where the abnormal crowd suddenly increases in the abnormal geographic grid by comparing the abnormal grid abnormal judgment value of the abnormal geographic grid with the number of people gathered in the abnormal geographic grid at the current moment, as the final abnormal geographic grid. By further performing statistical analysis on the geographic grid where the abnormal gathering occurs, the detection range of the abnormal crowd gathering detection is effectively narrowed, thereby improving the accuracy of the abnormal crowd gathering detection, and thus realizing the detection of abnormal crowd gathering that is efficient, timely, cost-effective and has a wide coverage.

[0129] Generally speaking, the same actual place often includes multiple geographical grids that are continuous in multiple geographical locations. In order to better determine whether the actual place is abnormal, the abnormal geographical grids can be further aggregated, so as to guide relevant personnel to prevent abnormal crowd gathering from a macro perspective. Figure 4 , Figure 4 FIG. 1 is a flow chart of a method for detecting abnormal crowds according to another embodiment of the present application. Figure 4 As shown, the method includes:

[0130] S401, obtaining coordinate information of an abnormal geographic grid.

[0131] In the embodiment of the present application, the abnormal geographical grid may be a geographical grid with an abnormal number of people gathered based on the sequence data to be detected, or it may be an abnormal geographical grid with an abnormal increase in the number of people further screened out based on the geographical grid with an abnormal number of people gathered.

[0132] Among them, the abnormal crowd gathering area generated based on the preliminary screened abnormal geographic grid and the abnormal crowd gathering area generated based on the abnormal geographic grid with abnormal crowd increase further screened out based on the geographic grid with abnormal crowd gathering number, can both realize the detection of abnormal crowd gathering situation.

[0133] Furthermore, after determining the abnormal geographic grid, the encoded character string that can represent the geographical location of the abnormal geographic grid carried in the historical signaling sequence data and / or real-time signaling sequence data carried in the sequence data to be detected corresponding to the abnormal geographic grid can be obtained, and the encoded character string of the geographical location can be decoded and calculated to obtain the coordinate information of the geographical location corresponding to the abnormal geographic grid.

[0134] Optionally, a geohash algorithm may be used. The geohash algorithm encodes the longitude and latitude coordinate information of the geographic location by a binary method, encoding the two-dimensional longitude and latitude coordinates into a one-dimensional character string.

[0135] For example, the longitude and latitude coordinates of the center of a certain geographic grid among N geographic grids are set to (39.92, 116.39), and the latitude coordinates in the center coordinates of the above geographic grid are encoded by the geohash algorithm.

[0136] The latitude range (-90,90) is divided into two intervals, namely (-90,0) and (0,90). The latitude of the center position of the geographic grid is compared with the two obtained latitude intervals. If the latitude coordinate falls into the first latitude interval obtained by the binary division, it is encoded as 0. If the latitude coordinate falls into the second latitude interval obtained by the binary division, it is encoded as 1.

[0137] Furthermore, the latitude interval (0,90) in which it falls is divided into two parts again to obtain two latitude intervals (0,45) and (45,90). The latitude coordinate of the center of the geographic grid is compared with the two latitude intervals obtained by the second binary division. If it falls into the first latitude interval obtained, it is encoded as 0, and if it falls into the second latitude interval obtained, it is encoded as 1.

[0138] The latitude interval it falls into is further divided into two, and the latitude coordinates of the center of the geographic grid are further encoded based on the two latitude intervals obtained, and so on. According to the preset string precision, the encoded string of the latitude coordinates of the center of the geographic grid is obtained. For example, if the string precision is set to 8 bits, then when the number of times the latitude interval is divided into two reaches 8 times, the latitude interval is stopped from being divided into two, and the string of the latitude coordinates of the center of the geographic grid encoded at the current precision is 10111000.

[0139] Furthermore, the acquired geographical location character string of the final abnormal geographical grid is decoded to obtain the longitude and latitude coordinates of the geographical location corresponding to the character string, thereby determining the coordinate information corresponding to the final abnormal geographical grid.

[0140] S402, based on the coordinate information, performing geographic grid aggregation processing on the abnormal geographic grids to generate an abnormal crowd area.

[0141] Each of the N geographic grids may cover a regular-shaped area in the map, such as a rectangle.

[0142] In general, the geographical scope covered by a large gathering of people is unstable and uncontrollable. For example, in a large-scale entertainment event, the area occupied by participants can be much larger than the area covered by a single geographical grid in the embodiment of the present application. Therefore, it is necessary to further aggregate the final abnormal geographical grid, merge the abnormal geographical grids that meet the aggregation conditions, and treat the merged abnormal geographical grids as a whole to generate an abnormal gathering area of ​​the crowd.

[0143] In an embodiment of the present application, the center point coordinate information of the abnormal geographic grids can be mapped to a map, and pairwise comparisons can be performed based on the center point coordinates of the abnormal geographic grids on the map. For the abnormal geographic grids that meet the aggregation conditions, geographic grid aggregation processing can be performed to generate a complete area where an abnormal gathering of people has occurred.

[0144] For example, based on the example proposed in the above embodiment, it is assumed that each geographic grid can map a square geographic location on the map, and the final abnormal geographic grid is determined to be W1, W 22 ,W 38Based on the geographic location code string it carries, the coordinate information corresponding to the three abnormal geographic grids is obtained, and the abnormal geographic grids W1, W 22 ,W 38 The coordinate information of the abnormal geographic grid is mapped to the map, and the coordinates of the four points of each abnormal geographic grid are obtained for calculation to determine the abnormal geographic grid W1, W 22 ,W 38 The central coordinates of the two geographic grids. Assuming that each geographic grid can cover a 150m×150m geographic grid on the map, when the distance between the center points of two geographic grids on the map is less than 2×150m, it can be understood that the two geographic grids belong to the same location on the map, and the two abnormal geographic grids that meet this condition can be aggregated.

[0145] Based on this setting, the abnormal geographic grids W1,W 22 ,W 38 The distance between the center coordinates of the abnormal geographic grid W1,W 22 The distance between the center coordinates of the abnormal geographic grid W is less than 2×150m. 38 Compared with the other two abnormal geographic grids W1,W 22 If the distance between the center point coordinates of the abnormal geographic grid W1, W 22 Aggregation processing is performed to generate the abnormal crowd gathering area Z1, and the abnormal geographic grid W 38 The area with abnormal population gathering is Z2.

[0146] It should be particularly noted that the embodiment of the present application is aimed at detecting crowd gatherings in the case of an abnormal sudden increase in the number of people. Although the abnormal crowd gathering area generated based on the abnormal geographic grid obtained through preliminary screening can also reflect the abnormal gathering of the crowd, it contains other abnormal situations besides the abnormal sudden increase in the number of people.

[0147] The crowd anomaly detection method proposed in the present application, through geographic grid aggregation processing of abnormal geographic grids, merges the abnormal geographic grids that meet the aggregation processing conditions to obtain the precise area of ​​abnormal crowd gathering, so that the results of abnormal crowd gathering detection can correspond to the buildings and / or scenes in the actual space, providing data support for the subsequent judgment of the abnormal crowd gathering situation and the execution of treatment measures.

[0148] Furthermore, it can be combined with Figure 5 understand, Figure 5 FIG. 1 is a flow chart of a method for detecting abnormal crowds according to another embodiment of the present application. Figure 5 As shown, the method includes:

[0149] S501, determining the coordinates of the center point of the abnormal geographic grid according to the coordinate information of the abnormal geographic grid.

[0150] According to the character string encoded with the geographic location coordinate information carried in the historical signaling sequence data and / or the real-time signaling sequence data in the sequence data to be detected of the abnormal geographic grid, the coordinate information corresponding to the geographic grid is obtained. By mapping the obtained coordinate information of the abnormal geographic grid to the map, the geographic location information to which the coverage range of the abnormal geographic grid on the map belongs can be obtained.

[0151] The geographic grid that can represent location information proposed in the embodiment of the present application is a geographic grid with a regular shape, such as a rectangle, etc. Generally, after determining the position and coordinate information of each geographic grid mapped to the map, the geometric center point corresponding to the geographic grid can be obtained, and the coordinates of the geometric center point mapped to the map are determined as the center point coordinates.

[0152] S502: Determine the distance between any two abnormal geographic grids based on the center point coordinates.

[0153] After determining the coordinates of the center points of the abnormal geographic grids, the distance between the two center points can be obtained based on the coordinates of the center points of every two abnormal geographic grids, thereby determining the distance between the two abnormal geographic grids.

[0154] Furthermore, the distance between the center coordinates of every two abnormal geographic grids is obtained until the distance between any two abnormal geographic grids in all the abnormal geographic grids is obtained.

[0155] For example, set the abnormal geographic grid to W1,W 22 ,W 38 , determine the distance between any two abnormal geographic grids is to obtain W1, W 22 The distance between 22 ,W 38 The distance between W1,W 38 The distance between.

[0156] S503: Group the abnormal geographic grids according to the distance and the coordinate information of the abnormal geographic grids to form an aggregated geographic grid group.

[0157] After obtaining the distance between any two geographic grids in all abnormal geographic grids, the abnormal geographic grids are grouped according to the distance between any two geographic grids to form an aggregated geographic grid group.

[0158] Furthermore, an abnormal geographic grid pair whose distance is less than a preset distance threshold may be selected, wherein any two abnormal geographic grids constitute an abnormal geographic grid pair.

[0159] A distance threshold between abnormal geographic grids can be set, any two abnormal geographic grids are combined into an abnormal geographic grid pair, and the distance between the two abnormal geographic grids included in each abnormal geographic grid pair is obtained.

[0160] The distance between the two abnormal geographic grids in each acquired abnormal geographic grid pair is compared with the set distance threshold. When the distance between the two abnormal geographic grids is less than the distance threshold, it can be understood that the geographical locations of the two abnormal geographic grids in the map are close and can be regarded as the same area, or the buildings and / or sites at the geographical locations mapped by the two abnormal geographic grids in the map are the same building and / or site.

[0161] Furthermore, an abnormal geographic grid pair is obtained, in which the distance between the two abnormal geographic grids is less than a preset distance threshold.

[0162] Furthermore, based on the coordinate information of the abnormal geographic grids, one or more spatially continuous aggregated geographic grid groups are determined from the abnormal geographic grid pairs, wherein the aggregated geographic grid group includes one or more abnormal geographic grid pairs.

[0163] After obtaining the abnormal geographic grid pairs that meet the aggregation conditions, the abnormal geographic grids belonging to the same building and / or site are further identified according to the coordinate information of each abnormal geographic grid in the abnormal geographic grid pairs.

[0164] Generally, a building and / or site at a certain geographical location has a certain area, and the area may contain one or more geographical locations mapped on the map by geographical grids. After obtaining the abnormal geographical grids in the abnormal geographical grid pairs that meet the aggregation conditions, the corresponding center point coordinates may be obtained, and the obtained center point coordinates may be mapped into the map. The abnormal geographical grid pairs that belong to the same building and / or site in geographical location may be screened out according to the location of the center point coordinates. Generally, in the dimension of spatial location, it is determined that the abnormal geographical grid pairs that belong to the same building and / or site are continuous in spatial location, so the abnormal geographical grid pairs that belong to the same building and / or site in geographical location selected above may be aggregated to generate an aggregated geographical grid group.

[0165] For example, the distance threshold between the center point coordinates of the abnormal geographic grids in the area with abnormal crowd gathering is set to L = 2 × 150m, and the coverage area of ​​each geographic grid mapped to the map is 150m × 150m. Any two abnormal geographic grids are grouped to generate abnormal geographic grid pairs.

[0166] Furthermore, the center point of each abnormal geographic grid in the acquired abnormal geographic grid pair is mapped to the map, and clustering is performed according to the comparison result between the distance between the center point coordinates and the distance threshold L. The center point distance between two abnormal geographic grids in any abnormal geographic grid pair is cyclically compared with the distance threshold L. When the distance between the center point coordinates of two abnormal geographic grids in a certain abnormal geographic grid pair is less than L, the two abnormal geographic grids are classified as the same area, until the distance between the center coordinates of all abnormal geographic grids is greater than or equal to the distance threshold L, the clustering is stopped, and then one or more abnormal geographic grid pairs are generated to generate aggregated geographic grid groups.

[0167] It should be noted that one or more abnormal geographic grid pairs belonging to the same building and / or site can generate an aggregated geographic grid group. One or more abnormal geographic grid pairs in the embodiments of the present application may belong to different buildings and / or sites. Therefore, the aggregated geographic grid group generated by the abnormal geographic grid pairs that meet the aggregation conditions can be one or more.

[0168] The method for detecting abnormal crowd gathering proposed in the present application obtains abnormal geographic grid grouping composed of abnormal geographic grid pairs continuous in spatial position through the central coordinate information of the abnormal geographic grid, which provides support for the subsequent generation of abnormal crowd gathering areas, so that abnormal crowd gathering detection can be realized efficiently, timely, cost-effective and with wide coverage.

[0169] Furthermore, it can be combined with Figure 6 understand, Figure 6 FIG. 1 is a flow chart of a method for detecting abnormal crowds according to another embodiment of the present application. Figure 6 As shown, the approach includes:

[0170] S601, obtaining the maximum coordinate and the minimum coordinate in the abnormal crowd area as the area boundary of the abnormal crowd area.

[0171] In order to further obtain the crowd density in the abnormal crowd gathering area and then accurately judge the degree of abnormal crowd gathering, it is necessary to obtain the regional boundary of the abnormal crowd gathering area.

[0172] After generating an abnormal crowd gathering area based on the aggregation and classification of abnormal geographic grids, the maximum coordinates and minimum coordinates of the abnormal crowd gathering area can be obtained according to the coordinate information corresponding to the abnormal geographic grids contained in the abnormal crowd gathering area, and the boundary line of the abnormal crowd gathering area can be generated based on the maximum coordinates and the minimum coordinates.

[0173] For example, based on the example proposed in step S402 above, the abnormal geographic grids W1, W 22The abnormal crowd gathering area Z1 and abnormal geographic grid W are generated. 38 The abnormal crowd gathering area Z2 is generated. By obtaining the abnormal geographic grid W1 and W 22 By obtaining the maximum longitude and minimum longitude and latitude of the abnormal geographic grid, the maximum coordinates and minimum coordinates of the abnormal crowd gathering area Z1 can be obtained. Similarly, by obtaining the maximum longitude and minimum longitude and latitude of the abnormal geographic grid, the maximum coordinates and minimum coordinates of the abnormal crowd gathering area Z2 can be obtained.

[0174] Furthermore, the maximum longitude and minimum longitude of the abnormal crowd gathering area Z1 are connected along the latitude line where the maximum latitude and the minimum latitude are located, based on the minimum distance between the two longitudes, and then the maximum latitude and the minimum latitude are connected along the longitude line where the maximum longitude and the minimum longitude are located, based on the minimum distance between the two latitudes, thereby generating the regional boundary of the abnormal crowd gathering area Z1. Correspondingly, the regional boundary of the abnormal crowd gathering area Z2 can be obtained by the same method.

[0175] S602, obtaining geographical grids belonging to the abnormal crowd area, and counting the number of people in each geographical grid belonging to the abnormal crowd area at the current moment and the number of people in history for a set period of time.

[0176] Generally speaking, the number of people involved in an abnormal crowd gathering can directly reflect the severity of the abnormal crowd gathering.

[0177] After determining the regional boundary of the abnormal crowd gathering area, all geographic grids within its coverage can be obtained, and then the real-time signaling sequence data corresponding to each geographic grid can be obtained. The real-time signaling sequence data corresponding to each geographic grid carries the number of people gathered in the geographic grid at the current moment. The number of people gathered at the current moment corresponding to all geographic grids contained in the abnormal crowd gathering area is added together to obtain the number of people gathered in the abnormal crowd gathering area at the current moment.

[0178] Accordingly, the historical signaling sequence data corresponding to each geographic grid in all geographic grids covered by the abnormal crowd gathering area can be obtained at the same time. The historical signaling sequence data is the historical crowd gathering situation of the geographic grid within the collection setting time range, which carries the historical crowd gathering number of the geographic grid within the collection setting time range corresponding to each geographic grid. Further, by adding the corresponding historical crowd gathering number of each geographic grid within the set time range, the historical crowd gathering number of the set time range in the abnormal crowd gathering area can be obtained.

[0179] S603, concatenating the current number of people with the historical number of people to generate number sequence data of the abnormal crowd area.

[0180] In order to more intuitively compare the crowd gathering situation at the current moment with the historical crowd gathering situation, in an embodiment of the present application, after obtaining the number of abnormal crowd gatherings at the current moment and the number of historical crowd gatherings for a set period of time, it is necessary to splice the number of abnormal crowd gatherings at the current moment and the number of historical crowd gatherings for a set period of time to generate number sequence data of the abnormal crowd gathering area.

[0181] For example, based on the above example, 50 geographic grids are tested for abnormal crowd gathering, and the abnormal crowd gathering area is determined to be Z1. The number of people gathered in the abnormal crowd gathering area Z1 at the current moment is counted, and the number of people gathered in each geographic grid covered by the area in the past 30 days is counted. After splicing the number of people gathered at the current moment with the number of people gathered in the history, the number of people in the abnormal crowd gathering area Z1 is generated.

[0182] Furthermore, the headcount series data can be used as the output result of the abnormal crowd aggregation detection in 50 geographic grids in this round.

[0183] The crowd anomaly detection method proposed in the present application generates the boundary of the abnormal crowd gathering area by obtaining the maximum coordinate and the minimum coordinate in the abnormal crowd area, so that the monitor can more accurately lock the scope of the abnormal crowd gathering, and generate number sequence data by splicing the number of abnormal crowd gatherings at the current moment and the number of historical crowd gatherings for a set period of time, so that the monitor can accurately judge the degree of gathering in the abnormal crowd gathering area at the current moment, and then propose effective treatment measures.

[0184] In order to better understand the above embodiments, Figure 7 , Figure 7 This is a flow chart of a method for detecting abnormal crowds according to another embodiment of the present application. Figure 7 As shown:

[0185] Abnormal crowd gathering is detected for N geographic grids, and historical signaling data corresponding to the N geographic grids are segmented according to preset time intervals to obtain signaling subsequence data corresponding to each geographic grid. The obtained signaling subsequence data are classified according to time type, and the signal subsequences under each time type are spliced ​​in chronological order, thereby generating historical signaling sequence data corresponding to multiple time types contained in each geographic grid.

[0186] Furthermore, the real-time signaling sequence data of N geographic grids at the current moment is obtained, and the corresponding historical signaling sequence data is obtained based on the time type to which the real-time signaling sequence data belongs. The real-time signaling sequence data is spliced ​​to the historical signaling sequence data to generate the sequence data to be detected, and the abnormal crowd gathering detection model is input to obtain the abnormal geographic grid where the abnormal crowd gathering occurs. Further, the abnormal geographic grids obtained are statistically analyzed to screen out the abnormal geographic grids where the crowd suddenly increases and determine them as the final abnormal geographic grids. The final abnormal geographic grids are clustered to generate an abnormal crowd gathering area, and the regional boundary of the abnormal crowd gathering area is obtained according to the maximum coordinates and the small coordinates of the abnormal crowd gathering area, and then the current moment of the crowd gathering number of all geographic grids in the abnormal crowd gathering area and the historical gathering number of the set time are obtained, and the two are spliced ​​to generate the number sequence data, and the number sequence data is output as the final detection result.

[0187] The crowd abnormality detection method proposed in the present application divides the geographic location information into N geographic grids, thereby reducing the difficulty of detecting abnormal crowd gatherings. Through signaling data, it is possible to determine whether an abnormal crowd gathering occurs, thereby improving the detection efficiency and achieving high-coverage abnormal crowd gathering detection. It is independent of video surveillance equipment, thus reducing the cost of detection, thereby achieving efficient, timely, cost-effective and wide-coverage detection of abnormal crowd gatherings.

[0188] Corresponding to the crowd anomaly detection methods proposed in the above-mentioned embodiments, an embodiment of the present application also proposes a crowd anomaly detection device. Since the crowd anomaly detection device proposed in the embodiment of the present application corresponds to the crowd anomaly detection methods proposed in the above-mentioned embodiments, the implementation methods of the above-mentioned crowd anomaly detection methods are also applicable to the crowd anomaly detection device proposed in the embodiment of the present application, and will not be described in detail in the following embodiments.

[0189] Figure 8 Schematic diagram of the structure of a crowd anomaly detection device according to an embodiment of the present application. Figure 8 As shown, the group anomaly detection device 100 includes an acquisition module 11, a splicing module 12, and a detection module 13, wherein:

[0190] An acquisition module 11 is used to acquire real-time signaling sequence data of terminal devices of N geographic grids at the current moment;

[0191] The acquisition module 11 is further used to determine, for any geographic grid among the N geographic grids, the historical signaling sequence data corresponding to the real-time signaling sequence data of any geographic grid according to the moment type of the current moment;

[0192] A splicing module 12, used for splicing the real-time signaling sequence data with the historical signaling sequence data to generate sequence data to be detected;

[0193] The detection module 13 is used to determine an abnormal geographical grid with abnormal population from N geographical grids based on the sequence data to be detected of each geographical grid.

[0194] Fig. 9 Schematic diagram of the structure of a crowd anomaly detection device according to an embodiment of the present application. Fig. 9 As shown, the crowd anomaly detection device 200 includes an acquisition module 21, a splicing module 22, a detection module 23, a calculation module 24, an abnormal geographic grid determination module 25, a coordinate acquisition module 26, an aggregation module 27, a segmentation module 28, an identification module 29, a boundary generation module 210, and a number acquisition module 211, wherein:

[0195] It should be noted that the acquisition module 11 , the splicing module 12 , the detection module 13 and the acquisition module 21 , the splicing module 22 , the detection module 23 have the same structure and function.

[0196] In the embodiment of the present application, the crowd anomaly detection device 200 further includes:

[0197] The calculation module 24 is used to obtain the mean and variance of the sequence data to be detected of the abnormal geographic grid, and determine the geographic grid abnormality judgment value based on the mean and variance;

[0198] The abnormal geographic grid determination module 25 is used to obtain the number of people in the abnormal geographic grid at the current moment, and select the abnormal geographic grid whose number of people at the current moment is greater than the abnormal determination value as the final abnormal geographic grid.

[0199] In the embodiment of the present application, the crowd anomaly detection device 200 further includes:

[0200] A coordinate acquisition module 26, used to acquire coordinate information of an abnormal geographic grid;

[0201] The aggregation module 27 is used to perform geographic grid aggregation processing on abnormal geographic grids based on coordinate information to generate abnormal crowd areas.

[0202] In the embodiment of the present application, the aggregation module 27 is also used to: determine the center point coordinates of the abnormal geographic grid based on the coordinate information of the abnormal geographic grid; determine the distance between any two abnormal geographic grids based on the center point coordinates; group the abnormal geographic grids according to the distance and the coordinate information of the abnormal geographic grid to form an aggregated geographic grid group; perform geographic grid aggregation processing on the abnormal geographic grids belonging to the same group to generate an abnormal population area.

[0203] In the embodiment of the present application, the aggregation module 27 is also used to: select abnormal geographic grid pairs whose distance is less than a preset distance threshold, wherein any two abnormal geographic grids constitute an abnormal geographic grid pair; based on the coordinate information of the abnormal geographic grids, determine one or more continuous aggregated geographic grid groups in spatial position from the abnormal geographic grid pairs, wherein the aggregated geographic grid group includes one or more abnormal geographic grid pairs.

[0204] In the embodiment of the present application, the crowd anomaly detection device 200 further includes:

[0205] The acquisition module 21 is also used to acquire historical signaling data of N geographic grids;

[0206] A segmentation module 28 is used to segment the historical signaling data of any geographic grid among the N geographic grids according to a preset time interval to generate a plurality of signaling subsequence data;

[0207] The splicing module 22 is also used to identify the time type of the signaling subsequence data, and splice the signaling subsequence data belonging to the same time category in chronological order from early to late, so as to obtain historical signaling sequence data of each time type.

[0208] In the embodiment of the present application, the crowd anomaly detection device 200 further includes:

[0209] The acquisition module 21 is further used to obtain the collection time length corresponding to the historical signaling data in any geographical grid;

[0210] An identification module 29, configured to perform time type identification based on a clustering algorithm in response to the acquisition time length being greater than a preset time length;

[0211] The identification module 29 is further configured to perform time type identification according to the classification rule in response to the acquisition time length being less than or equal to the preset time length.

[0212] In the embodiment of the present application, the crowd anomaly detection device 200 further includes:

[0213] The boundary generation module 210 is used to obtain the maximum coordinate and the minimum coordinate in the abnormal crowd area as the area boundary of the abnormal crowd area.

[0214] In the embodiment of the present application, the crowd anomaly detection device 200 further includes:

[0215] The number acquisition module 211 is used to acquire the geographical grids belonging to the abnormal crowd area, and count the number of people in each geographical grid belonging to the abnormal crowd area at the current moment, as well as the historical number of people for a set period of time;

[0216] The splicing module 22 is also used to splice the number of people at the current moment with the number of people in history to generate the number sequence data of the abnormal crowd area.

[0217] The crowd anomaly detection device proposed in the present application obtains historical signaling data of N geographic grids, and performs segmentation and classification to generate historical signaling sequence data corresponding to multiple time types included in each geographic grid. The real-time signaling sequence data of N geographic grids at the current moment is obtained, and the corresponding historical signaling sequence data is obtained based on the time type to which the real-time signaling sequence data belongs. The two are spliced ​​to generate sequence data to be detected, and after inputting into the crowd abnormal gathering detection model, the abnormal geographic grid where the crowd abnormally gathers is obtained. Further, the obtained abnormal geographic grid is statistically analyzed, and the geographic grid with a sudden increase in the crowd and abnormal gathering is screened out to determine as the final abnormal geographic grid. Furthermore, the determined abnormal geographic grids are classified to generate abnormal crowd gathering areas, and the final detection results are output. The crowd abnormality detection method proposed in the present application divides the geographic location information into N geographic grids, thereby reducing the difficulty of detecting abnormal crowd gatherings. Through signaling data, it is possible to determine whether an abnormal crowd gathering occurs, thereby improving the detection efficiency and achieving high-coverage abnormal crowd gathering detection. It is independent of video surveillance equipment, thus reducing the cost of detection, thereby achieving efficient, timely, cost-effective and wide-coverage detection of abnormal crowd gatherings.

[0218] To achieve the above embodiments, the present application also proposes an electronic device, a computer-readable storage medium and a computer program product.

[0219] Fig.10 A schematic block diagram of an example electronic device 1000 that can be used to implement an embodiment of the present application is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workbenches, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present application described and / or required herein.

[0220] like Fig.10 As shown, the device 1000 includes a memory 101, a processor 102, and a computer program stored in the memory 101 and executable on the processor 102. When the processor 102 executes the program instructions, the crowd anomaly detection method proposed in the above embodiment is implemented.

[0221] The electronic device proposed in the embodiment of the present application executes a computer program stored in the memory 101 through the processor 102, detects abnormal crowd gathering for N geographic grids, divides the historical signaling data corresponding to the N geographic grids according to preset time intervals, obtains signaling subsequence data corresponding to each geographic grid, classifies the obtained signaling subsequence data according to time type, and splices the signaling subsequences under each time type in chronological order, thereby generating historical signaling sequence data corresponding to multiple time types contained in each geographic grid.

[0222] Furthermore, the real-time signaling sequence data of N geographic grids at the current moment is obtained, and the corresponding historical signaling sequence data is obtained based on the time type to which the real-time signaling sequence data belongs. The real-time signaling sequence data is spliced ​​to the historical signaling sequence data to generate the sequence data to be detected, and the abnormal crowd gathering detection model is input to obtain the abnormal geographic grid where the abnormal crowd gathering occurs.

[0223] Furthermore, the acquired abnormal geographic grids are statistically analyzed to screen out abnormal geographic grids with sudden increase in crowds and determine them as the final abnormal geographic grids. The final abnormal geographic grids are clustered to generate abnormal crowd gathering areas. According to the maximum coordinates and small coordinates of the abnormal crowd gathering areas, the regional boundaries of the abnormal crowd gathering areas are obtained, and then the number of people gathered at the current moment and the historical number of people gathered for a set period of time for all geographic grids in the abnormal crowd gathering areas are obtained. The two are spliced ​​to generate population sequence data, and the population sequence data is output as the final detection result.

[0224] The crowd abnormality detection method proposed in the present application divides the geographic location information into N geographic grids, thereby reducing the difficulty of detecting abnormal crowd gatherings. Through signaling data, it is possible to determine whether an abnormal crowd gathering occurs, thereby improving the detection efficiency and achieving high-coverage abnormal crowd gathering detection. It is independent of video surveillance equipment, thus reducing the cost of detection, thereby achieving efficient, timely, cost-effective and wide-coverage detection of abnormal crowd gatherings.

[0225] The embodiment of the present application proposes a computer-readable storage medium, on which a computer program is stored. When the program is executed by the processor 102, the crowd anomaly detection method proposed in the above embodiment is implemented.

[0226] The computer-readable storage medium of the embodiment of the present application stores a computer program and is executed by a processor. The processor 102 executes the computer program stored in the memory 101 to detect abnormal crowd gathering for N geographic grids, divide the historical signaling data corresponding to the N geographic grids according to preset time intervals, obtain signaling subsequence data corresponding to each geographic grid, classify the obtained signaling subsequence data according to time type, and splice the signaling subsequences under each time type in chronological order, thereby generating historical signaling sequence data corresponding to multiple time types contained in each geographic grid.

[0227] Furthermore, the real-time signaling sequence data of N geographic grids at the current moment is obtained, and the corresponding historical signaling sequence data is obtained based on the time type to which the real-time signaling sequence data belongs. The real-time signaling sequence data is spliced ​​to the historical signaling sequence data to generate the sequence data to be detected, and the abnormal crowd gathering detection model is input to obtain the abnormal geographic grid where the abnormal crowd gathering occurs.

[0228] Furthermore, the acquired abnormal geographic grids are statistically analyzed to screen out abnormal geographic grids with sudden increase in crowds and determine them as the final abnormal geographic grids. The final abnormal geographic grids are clustered to generate abnormal crowd gathering areas. According to the maximum coordinates and small coordinates of the abnormal crowd gathering areas, the regional boundaries of the abnormal crowd gathering areas are obtained, and then the number of people gathered at the current moment and the historical number of people gathered for a set period of time for all geographic grids in the abnormal crowd gathering areas are obtained. The two are spliced ​​to generate population sequence data, and the population sequence data is output as the final detection result.

[0229] The crowd abnormality detection method proposed in the present application divides the geographic location information into N geographic grids, thereby reducing the difficulty of detecting abnormal crowd gatherings. Through signaling data, it is possible to determine whether an abnormal crowd gathering occurs, thereby improving the detection efficiency and achieving high-coverage abnormal crowd gathering detection. It is independent of video surveillance equipment, thus reducing the cost of detection, thereby achieving efficient, timely, cost-effective and wide-coverage detection of abnormal crowd gatherings.

[0230] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations may include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0231] The program code for implementing the method itself can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code can be executed entirely on the machine, partially on the machine, partially on the machine as a stand-alone software package and partially on a remote machine, or entirely on a remote machine or server.

[0232] In the context of the present application, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0233] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0234] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a geo-grid browser through which a user can interact with an implementation of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communications geo-grid). Examples of communications geo-grids include: a local area network (LAN), a wide area network (WAN), the Internet, and a blockchain geo-grid.

[0235] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication geographic grid. The relationship between the client and the server is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services ("Virtual Private Server", or "VPS" for short). The server may also be a server for a distributed system, or a server combined with a blockchain.

[0236] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0237] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of this application, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0238] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or more executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.

[0239] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute the instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways if necessary, and then stored in a computer memory.

[0240] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0241] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0242] In addition, each functional unit in each embodiment of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0243] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present application. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present application.

[0244] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this application can be executed in parallel, sequentially or in different orders, as long as the expected results of the technical solution disclosed in this application can be achieved, and this document is not limited here.

[0245] The above specific implementations do not constitute a limitation on the protection scope of this application. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of this application should be included in the protection scope of this application.

Claims

1. A method for detecting abnormal crowds, characterized in that: include: Obtain real-time signaling sequence data of terminal devices in N geographic grids at the current moment; For any geographical grid among the N geographical grids, determining the historical signaling sequence data corresponding to the real-time signaling sequence data of the any geographical grid according to the moment type of the current moment; splicing the real-time signaling sequence data with the historical signaling sequence data to generate sequence data to be detected; Based on the sequence data to be detected of each geographic grid, determining an abnormal geographic grid with abnormal population from the N geographic grids; Obtaining coordinate information of the abnormal geographic grid; Determine the coordinates of the center point of the abnormal geographic grid according to the coordinate information of the abnormal geographic grid; Based on the center point coordinates, determining the distance between any two abnormal geographic grids; According to the distance and the coordinate information of the abnormal geographic grid, the abnormal geographic grid is grouped to form an aggregated geographic grid group; The abnormal geographic grids belonging to the same group are aggregated to generate abnormal population areas.

2. The method according to claim 1, characterized in that After determining the abnormal geographical grid with abnormal population from the N geographical grids, the method further includes: Obtaining the mean and variance of the sequence data to be detected of the abnormal geographic grid, and determining the geographic grid abnormality determination value based on the mean and the variance; The number of people in the abnormal geographic grid at the current moment is obtained, and the abnormal geographic grid whose number of people at the current moment is greater than the abnormal determination value is selected as the final abnormal geographic grid.

3. The method according to claim 1, characterized in that The determining of an abnormal geographical grid having an abnormal population from the N geographical grids based on the sequence data to be detected of each geographical grid further includes: A time series anomaly detection model is trained, and based on the trained time series anomaly detection model, an abnormal geographical grid with abnormal population among the N geographical grids is determined.

4. The method according to claim 1, characterized in that The step of grouping the abnormal geographic grids according to the distance and the coordinate information of the abnormal geographic grids to form an aggregated geographic grid group includes: Selecting an abnormal geographic grid pair whose distance is less than a preset distance threshold, wherein any two abnormal geographic grids constitute an abnormal geographic grid pair; Based on the coordinate information of the abnormal geographic grids, one or more spatially continuous aggregated geographic grid groups are determined from the abnormal geographic grid pairs, wherein the aggregated geographic grid group includes one or more abnormal geographic grid pairs.

5. The method according to claim 1, characterized in that Before acquiring the real-time signaling sequence data of the terminal devices of the N geographical grids at the current moment, the method further includes: Acquire historical signaling data of the N geographic grids; For any geographical grid among the N geographical grids, data segmentation is performed on the historical signaling data of the any geographical grid according to a preset time interval to generate a plurality of signaling subsequence data; The time type of the signaling subsequence data is identified, and the signaling subsequence data belonging to the same time category are spliced ​​in chronological order from early to late to obtain historical signaling sequence data of each time type.

6. The method according to claim 4, characterized in that Also includes: Obtaining the collection time length corresponding to the historical signaling data in any geographic grid; In response to the acquisition time length being greater than a preset time length, performing time type recognition based on a clustering algorithm; In response to the acquisition time length being less than or equal to the preset time length, time type identification is performed according to a classification rule.

7. The method according to claim 1, characterized in that After generating the abnormal crowd area, the method further includes: The maximum coordinate and the minimum coordinate in the abnormal crowd area are obtained as the area boundary of the abnormal crowd area.

8. The method according to claim 1, characterized in that After generating the abnormal crowd area, the method further includes: Obtaining geographic grids belonging to the abnormal crowd area, and counting the number of people in each geographic grid belonging to the abnormal crowd area at the current moment, as well as the historical number of people for a set period of time; The number of people at the current moment and the number of people in history are concatenated to generate sequence data of the number of people in the abnormal crowd area.

9. A crowd anomaly detection device, characterized in that: include: An acquisition module, used to acquire real-time signaling sequence data of terminal devices of N geographic grids at the current moment; The acquisition module is further used to determine, for any geographic grid among the N geographic grids, the historical signaling sequence data corresponding to the real-time signaling sequence data of any geographic grid according to the moment type of the current moment; A splicing module, used for splicing the real-time signaling sequence data with the historical signaling sequence data to generate sequence data to be detected; A detection module, configured to determine an abnormal geographical grid having an abnormal population from among the N geographical grids based on the sequence data to be detected of each geographical grid; A coordinate acquisition module, used to acquire the coordinate information of the abnormal geographic grid; An aggregation module, used to determine the coordinates of the center point of the abnormal geographic grid according to the coordinate information of the abnormal geographic grid; Based on the center point coordinates, determining the distance between any two abnormal geographic grids; According to the distance and the coordinate information of the abnormal geographic grid, the abnormal geographic grid is grouped to form an aggregated geographic grid group; The abnormal geographic grids belonging to the same group are aggregated to generate abnormal population areas.

10. The device according to claim 9, characterized in that Also includes: A calculation module, used to obtain the mean and variance of the sequence data to be detected of the abnormal geographic grid, and determine the geographic grid abnormality judgment value based on the mean and the variance; The abnormal geographic grid determination module is used to obtain the number of people in the abnormal geographic grid at the current moment, and select the abnormal geographic grid whose number of people at the current moment is greater than the abnormal determination value as the final abnormal geographic grid.

11. The device according to claim 9, characterized in that The aggregation module is further used for: Selecting an abnormal geographic grid pair whose distance is less than a preset distance threshold, wherein any two abnormal geographic grids constitute an abnormal geographic grid pair; Based on the coordinate information of the abnormal geographic grids, one or more spatially continuous aggregated geographic grid groups are determined from the abnormal geographic grid pairs, wherein the aggregated geographic grid group includes one or more abnormal geographic grid pairs.

12. The device according to claim 9, characterized in that Also includes: The acquisition module is further used to acquire the historical signaling data of the N geographic grids; A segmentation module, configured to segment the historical signaling data of any one of the N geographical grids according to a preset time interval to generate a plurality of signaling subsequence data; The splicing module is also used to identify the time type of the signaling subsequence data, and splice the signaling subsequence data belonging to the same time category in chronological order from early to late, so as to obtain historical signaling sequence data of each time type.

13. The device according to claim 11, characterized in that Also includes: The acquisition module is further used to acquire the collection time length corresponding to the historical signaling data in any geographic grid; an identification module, configured to perform time type identification based on a clustering algorithm in response to the acquisition time length being greater than a preset time length; The identification module is further configured to perform time type identification according to a classification rule in response to the acquisition time length being less than or equal to the preset time length.

14. The device according to claim 9, characterized in that Also includes: The boundary generation module is used to obtain the maximum coordinate and the minimum coordinate in the abnormal crowd area as the area boundary of the abnormal crowd area.

15. The device according to claim 9, characterized in that Also includes: A number acquisition module is used to acquire the geographical grids belonging to the abnormal crowd area, and count the number of people in each geographical grid belonging to the abnormal crowd area at the current moment, as well as the historical number of people for a set period of time; The splicing module is also used to splice the number of people at the current moment with the historical number of people to generate the number sequence data of the abnormal crowd area.

16. An electronic device, comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 8.

17. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-8.

18. A computer program product, comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 8.

Images