Non-transitory computer-readable storage medium, information processing method

Abstract

To facilitate appropriate determination of occurrence of dangerous congestion.SOLUTION: The information processing apparatus 100 acquires a plurality of pieces of position data 101 indicating positions of respective persons in a target region at different points in time. The information processing device 100 generates a diagram 110 indicating a timing at which each shape of one or more shapes of different types appears and disappears according to a change in resolution on the basis of the plurality of pieces of acquired position data 101 by topological data analysis. The information processing device 100 generates the first feature quantity data 111 corresponding to the generated diagram 110. The information processing apparatus 100 learns, based on the generated first feature value data 111, a model 120 having a function of outputting a result of analyzing a risk due to a congestion situation in a target area according to input feature value data representing a spatial feature in the target area.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an information processing program, an information processing method, and an information processing device. [Background technology]

[0002] To prevent crowd accidents, there is a conventional technology that determines that dangerous congestion that makes a crowd accident likely to occur has occurred when the density of people in a specific location exceeds a threshold, and issues a warning that the probability of a crowd accident occurring is above a certain level.

[0003] Prior art, for example, involves generating the degree of congestion within a railway vehicle based on images from a camera installed inside the vehicle, generating multiple types of congestion information with a hierarchical relationship based on the generated congestion degree, and transmitting the information according to the type of display unit. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2023-079602 Summary of the Invention [Problem to be solved by the invention]

[0005] However, with conventional technology, it is difficult to properly determine whether dangerous congestion that is likely to cause a crowd accident has occurred. For example, even if the flow of people is regular in a particular location and a crowd accident is unlikely to occur, if the density of people in the particular location exceeds a threshold, it may be erroneously determined that dangerous congestion that is likely to cause a crowd accident has occurred.

[0006] In one aspect, the present invention aims to facilitate appropriate determination of the occurrence of dangerous congestion. [Means for solving the problem]

[0007] According to one embodiment, an information processing program, an information processing method, and an information processing device are proposed that acquire a plurality of position data indicating the positions of each moving body in a target area at different times, generate first feature data representing spatial characteristics of the area based on the acquired plurality of position data through topological data analysis, the first feature data representing spatial characteristics of the area corresponding to a diagram showing the timing at which each of one or more shapes of different types that may be formed by combinations of the positions of the moving bodies in the area appear and disappear depending on changes in resolution, and learn a model that outputs results of an analysis of the risk due to congestion in the area based on the generated first feature data in accordance with the input feature data representing the spatial characteristics of the area. [Effects of the Invention]

[0008] According to one aspect, it is possible to facilitate appropriate determination of the occurrence of dangerous congestion. [Brief explanation of the drawings]

[0009] [Figure 1] FIG. 1 is an explanatory diagram illustrating an example of an information processing method according to an embodiment. [Figure 2] FIG. 2 is an explanatory diagram illustrating an example of an information processing system 200. As shown in FIG. [Figure 3] FIG. 3 is a block diagram showing an example of the hardware configuration of the information processing device 100. As shown in FIG. [Figure 4] FIG. 4 is a block diagram showing an example of the hardware configuration of the image capturing device 201. As shown in FIG. [Figure 5] FIG. 5 is a block diagram showing an example of the functional configuration of the information processing device 100. As shown in FIG. [Figure 6] FIG. 6 is an explanatory diagram (part 1) showing an example of the operation of the information processing device 100. [Figure 7] FIG. 7 is an explanatory diagram (part 2) showing an example of the operation of the information processing device 100. [Figure 8] FIG. 8 is an explanatory diagram (part 3) showing an example of the operation of the information processing device 100. [Figure 9]FIG. 9 is an explanatory diagram (part 4) showing an example of the operation of the information processing device 100. [Figure 10] FIG. 10 is an explanatory diagram (part 5) showing an example of the operation of the information processing device 100. [Figure 11] FIG. 11 is an explanatory diagram (part 6) showing an example of the operation of the information processing device 100. [Figure 12] FIG. 12 is an explanatory diagram (part 7) showing an example of the operation of the information processing device 100. [Figure 13] FIG. 13 is an explanatory diagram (part 1) showing a specific example of the operation of the information processing device 100. [Figure 14] FIG. 14 is an explanatory diagram (part 2) showing a specific example of the operation of the information processing device 100. [Figure 15] FIG. 15 is an explanatory diagram showing an example of the output screen. [Figure 16] FIG. 16 is a flowchart illustrating an example of the overall processing procedure. DETAILED DESCRIPTION OF THE INVENTION

[0010] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, an information processing program, an information processing method, and an information processing device according to embodiments of the present invention will be described in detail with reference to the accompanying drawings.

[0011] (An example of an information processing method according to an embodiment) 1 is an explanatory diagram illustrating an example of an information processing method according to an embodiment. The information processing device 100 is a computer that facilitates appropriate determination of the occurrence of dangerous congestion that is likely to cause a crowd accident. The information processing device 100 is, for example, a server or a PC (Personal Computer).

[0012] Congestion is a state in which the density, which indicates the number of people per unit area, exceeds a certain level. A crowd accident is an event that can occur in a specific location due to congestion caused by the presence of a narrow passage or obstacle, or by people gathering for an event, for example. In a crowd accident, people may pile up and fall, or be crushed, resulting in deaths or injuries. A specific location is, for example, a public space where people gather. A public space is, for example, a stadium, a station, or a busy shopping district.

[0013] Examples of crowd accidents include the Akashi Fireworks Festival Footbridge on July 21, 2001, Kangjulhan Stadium on October 1, 2022, and Itaewon on October 29, 2022.

[0014] For this reason, it is desirable to prevent crowd accidents. For example, in order to prevent crowd accidents, it is desirable to determine whether dangerous congestion that could lead to a crowd accident has occurred. Dangerous congestion is congestion in a state where a crowd accident is likely to occur. Dangerous congestion is congestion in a state where people's flow is stagnated, for example. When people's flow is stagnated, a distance that would normally take 10 minutes to travel can take several hours. Specifically, it is desirable to determine whether dangerous congestion that could lead to a crowd accident has occurred in public spaces, to quickly detect the occurrence of dangerous congestion that could lead to a crowd accident, and to take measures to address the dangerous congestion that has occurred.

[0015] Therefore, for example, a first method can be considered in which a dangerous congestion is determined to have occurred when the density of people in a specific location exceeds a first threshold. The density is, for example, the value obtained by dividing the number of people in the specific location by the area of ​​the specific location. For example, in the first method, when it is determined that a dangerous congestion has occurred, a warning indicating the occurrence of a dangerous congestion can be issued to a manager who manages safety in the specific location or a worker who physically manages the flow of people in the specific location. Here, for example, the following Reference 1 can be referred to for the first method.

[0016] Reference 1: Crowd Management Study Group, Social Collaboration Division, University of Tokyo (2020) "General Theory of Crowd Management - Theory and Practice" University of Tokyo Press

[0017] However, with the first method, it is difficult to appropriately determine whether dangerous congestion has occurred. For example, if a specific location is a relatively large space or has a relatively complex shape, the density of people in the specific location is not uniform, making it difficult to determine whether dangerous congestion has occurred. In fact, public spaces where people gather tend to be relatively large spaces or have relatively complex shapes. Specifically, even if dangerous congestion has occurred in some areas of the specific location, the density of people in the entire specific location may not exceed the first threshold, and it may be erroneously determined that dangerous congestion has not occurred.

[0018] In addition to dangerous congestion, there may also be safe congestion, which is a state in which crowd accidents are unlikely to occur. Safe congestion is congestion in which the flow of people is not impeded and is orderly. Safe congestion is congestion in which, for example, the flow of people is orderly, so that even if the density of people is relatively high, the probability of a crowd accident occurring does not exceed a certain level. Specific examples of safe congestion include congestion at train stations during normal rush hour.

[0019] In contrast, the first method described above is unable to distinguish between dangerous congestion and safe congestion and appropriately determine whether dangerous congestion has occurred. For example, the first method described above may erroneously determine that dangerous congestion has occurred if the density of people in a specific location exceeds a first threshold, even if the flow of people is regular in that location, dangerous congestion has not occurred, and a crowd accident is unlikely to occur.

[0020] Therefore, it is conceivable that warnings will be issued too frequently to the above-mentioned manager or worker. As a result, it is conceivable that the physical or psychological workload on the above-mentioned manager or worker will increase. It is also conceivable that the above-mentioned manager or worker will be induced to enter a psychological state in which they will disregard the warnings. This may make it difficult to prevent crowd accidents.

[0021] In addition to the first method, a second method can be considered in which the speed of people is calculated by tracking the movement of people in a specific location while identifying individuals based on images taken of the specific location, and if the calculated speed is equal to or less than a second threshold, it is determined that dangerous congestion has occurred. Tracking the movement of people based on images is called object tracking.

[0022] Even with the second method, it is difficult to accurately determine whether dangerous congestion has occurred. For example, it is difficult to accurately track the movements of people in a specific location while identifying individuals based on an image of the specific location, making it difficult to accurately determine whether dangerous congestion has occurred. Specifically, this can result in an increase in the processing time or processing load required to accurately track the movements of people.

[0023] Therefore, in this embodiment, an information processing method that can facilitate appropriate determination of the occurrence of dangerous congestion will be described. With this information processing method, for example, it is possible to distinguish between dangerous congestion and safe congestion and appropriately determine whether dangerous congestion has occurred in a target area.

[0024] In FIG. 1, there is a target area where it is determined whether dangerous congestion has occurred. The target area is an area where a plurality of moving objects exist. The moving objects are, for example, people. The target area is, for example, an area where a plurality of people exist. Specifically, the target area is a public space where people gather. The public space is, for example, a stadium, a station, or a busy shopping district. The information processing device 100 may store, for example, information that identifies the target area.

[0025] (1-1) The information processing device 100 acquires multiple pieces of position data 101. The multiple pieces of position data 101 indicate the positions of each person in a target area at different times. Here, the combination of people in the target area at a first time point and the combination of people in the target area at a second time point may be completely identical, may partially overlap, or may be completely different. The position data 101 indicates, for example, the position of each of one or more people in the target area without identifying individuals.

[0026] Here, the position data 101 preferably indicates the position of each of one or more people who actually exist in the target area. On the other hand, the position data 101 may be virtual data indicating the position of each of one or more people who are hypothesized to exist in the target area. In this case, the information processing device 100 acquires the position data 101 indicating the position of each person in the target area, specifically obtained by a simulator. In this case, the information processing device 100 specifically utilizes Agent Based Simulation.

[0027] (1-2) The information processing device 100 generates a diagram 110 based on the acquired multiple position data 101 through topological data analysis. The diagram 110 represents the timing at which one or more different types of shapes appear and disappear depending on the change in resolution. Each shape is a shape that can be formed by a combination of people's positions in the target area. The one or more shapes include, for example, at least any of a connected shape, a ring shape, and a hollow shape. A ring is a shape that exists on a two-dimensional plane. A hollow is, for example, a shape that exists in a three-dimensional space. The resolution indicates the size at which the person's position is treated. Specifically, the diagram 110 is a PD (Persistence Diagram).

[0028] The information processing device 100 generates first feature amount data 111 corresponding to the generated diagram 110. The first feature amount data 111 represents spatial features in a target area. The information processing device 100 generates a feature vector that serves as the first feature amount data 111 based on the diagram 110. Specifically, the first feature amount data 111 is a feature vector based on a persistence image (PI). Specifically, the first feature amount data 111 may be the PI itself.

[0029] (1-3) The information processing device 100 trains the model 120 based on the generated first feature amount data 111. The model 120 has a function of outputting an analysis result of the risk due to the congestion situation in the target area according to the input feature amount data representing the spatial characteristics of the target area. The model 120 is, for example, a neural network. The model 120 may also be a mathematical formula that classifies the presence or absence of risk due to the congestion situation into two clusters. The information processing device 100 trains the model 120 by supervised learning or unsupervised learning based on the generated first feature amount data 111. This allows the information processing device 100 to train the model 120 that can appropriately determine the occurrence of dangerous congestion.

[0030] The information processing device 100 can learn a model 120 that can distinguish between dangerous congestion and safe congestion and appropriately determine whether dangerous congestion has occurred. Specifically, the information processing device 100 can learn a model 120 that reflects spatial characteristics in the event of dangerous congestion and the tendency for the spatial characteristics to persist, via the diagram 110. Therefore, specifically, the information processing device 100 can learn a model 120 that determines whether dangerous congestion has occurred from a combination of people's positions in an area, without using the density of people in the target area.

[0031] Furthermore, the information processing device 100 can learn a model 120 that can appropriately determine whether dangerous congestion has occurred from only the positional information of people at each time, without identifying individuals or tracking the movements of people in a target area, for example, via the diagram 110. Therefore, the information processing device 100 can avoid an increase in processing time or processing load for tracking the movements of people in a target area when determining whether dangerous congestion has occurred, for example.

[0032] Here, the case where the functions of the information processing device 100 are realized by a single computer has been described, but this is not limiting. For example, the functions of the information processing device 100 may be realized by cooperation of multiple computers. For example, the functions of the information processing device 100 may be realized on the cloud.

[0033] Here, the case where the moving object is a person has been described, but this is not limiting. For example, the moving object may be a ship, a vehicle, an airplane, a drone, or the like. For example, if the moving object is a ship, it is conceivable that dangerous congestion of ships will occur in a target area such as a harbor with a relatively complex shape. In response to this, the information processing device 100 can appropriately determine whether dangerous congestion of ships has occurred.

[0034] (An example of the information processing system 200) Next, an example of an information processing system 200 to which the information processing device 100 shown in FIG. 1 is applied will be described with reference to FIG.

[0035] 2 is an explanatory diagram showing an example of an information processing system 200. In FIG. 2, the information processing system 200 includes an information processing device 100, one or more image capturing devices 201, and one or more client devices 202.

[0036] In the information processing system 200, the information processing device 100 and the imaging device 201 are connected via a wired or wireless network 210. The network 210 is, for example, a local area network (LAN), a wide area network (WAN), the Internet, etc. The information processing device 100 and the client device 202 are also connected via the wired or wireless network 210.

[0037] The imaging device 201 is a computer that captures an image of a target area for determining whether dangerous congestion has occurred. One or more imaging devices 201 are, for example, associated with different areas. The imaging device 201 captures an image of the area corresponding to the imaging device 201 at each time point, thereby generating an image of the area corresponding to the imaging device 201. The imaging device 201 analyzes the generated image at each time point, thereby generating position data indicating the position of each person in the area corresponding to the imaging device 201. The imaging device 201 transmits the generated position data to the information processing device 100 at each time point. The imaging device 201 is, for example, a fixed camera. The imaging device 201 may be, for example, a drone. The imaging device 201 may be, for example, a smartphone.

[0038] The information processing device 100 is a computer that facilitates appropriate determination of the occurrence of dangerous congestion in a specific area. The information processing device 100 stores a position data set that compiles, for each of one or more areas, a plurality of position data corresponding to different points in time within a plurality of periods. The position data indicates the position of each person in any of the areas. For example, the position data may indicate the position of each person in any of the areas without identifying the individual.

[0039] The information processing device 100 acquires the position data, for example, by receiving the position data from the imaging device 201. The information processing device 100 may also acquire the position data, for example, by accepting input of the position data. For example, based on the acquired position data, the information processing device 100 stores, for each of one or more areas, a position data set that compiles multiple pieces of position data corresponding to different points in time within each of multiple periods.

[0040] The information processing device 100 generates a PD for each period for each area based on a position data set for each period. The information processing device 100 generates a PI for each period for each area based on the PD generated for each period. The information processing device 100 generates a feature vector for each period for each area based on the PI generated for each period. The information processing device 100 learns a model that realizes a function of outputting the presence or absence of danger due to the congestion situation in the area based on the feature vector generated for each period for each area.

[0041] The information processing device 100 stores, for each region, a position data set that compiles together multiple pieces of position data corresponding to different points in time within a predetermined period. The information processing device 100 acquires the position data, for example, by receiving the position data from the imaging device 201. The information processing device 100 may also acquire the position data, for example, by accepting input of the position data. The information processing device 100 stores, for each region, a position data set that compiles together multiple pieces of position data corresponding to different points in time within a predetermined period, based on the acquired position data, for example.

[0042] The information processing device 100 generates a PD for a predetermined period for each region based on a position data set for the predetermined period. The information processing device 100 generates a PI for the predetermined period for each region based on the generated PD for the predetermined period. The information processing device 100 generates a feature vector for the predetermined period for each region based on the generated PI for the predetermined period.

[0043] The information processing device 100 generates and outputs information indicating whether or not there is a risk of congestion in each area during a predetermined period, using a trained model based on the generated feature vector for each area over a predetermined period. The output format may be, for example, display on a display, printout on a printer, transmission to another computer, or storage in a memory area.

[0044] The information processing device 100 transmits information indicating whether or not there is a risk due to congestion in each area during a predetermined period of time to another computer. The other computer is, for example, a client device 202. The information processing device 100 may output, for example, the presence or absence of a risk due to congestion in each area during a predetermined period of time so that the user can refer to it. The user is, for example, an administrator who manages safety in each area. The information processing device 100 is, for example, a server or a PC.

[0045] Client device 202 is a computer that outputs information indicating whether or not there is a risk due to congestion in each area over a predetermined period of time. Client device 202 is used, for example, by a manager who manages safety in each area or a worker who physically manages the flow of people in each area. Client device 202 receives information indicating whether or not there is a risk due to congestion in each area over a predetermined period of time from information processing device 100.

[0046] The client device 202 outputs information indicating whether or not there is a risk due to congestion in each area for a predetermined period of time so that the user can refer to it. For example, the client device 202 may output a warning indicating the risk due to congestion for an area among multiple areas that is at risk due to congestion so that the user can refer to it. This makes it easier for the client device 202 to manage safety in each area. The client device 202 is, for example, a PC, a tablet device, a smartphone, or a wearable device.

[0047] Here, the case where position data indicating the position of each person in the area is generated by analyzing an image of the area has been described, but this is not limiting. For example, the position data may be generated based on measurements from various sensors, such as an infrared sensor or a weight sensor, rather than an image. Furthermore, for example, the position data may be generated by detecting a device carried by each person in the area. The device may be, for example, a smartphone or a wearable device. The device may also be, for example, an IC tag.

[0048] Here, the case where the information processing device 100 is a device different from the imaging device 201 has been described, but this is not limiting. For example, the information processing device 100 may have the function of the imaging device 201 and operate as the imaging device 201. In this case, the information processing system 200 may not include the imaging device 201. Furthermore, the case where the information processing device 100 is a device different from the client device 202 has been described, but this is not limiting. For example, the information processing device 100 may have the function of the client device 202 and operate as the client device 202. In this case, the information processing system 200 may not include the client device 202.

[0049] (Application example of information processing system 200) Next, a description will be given of an application example of the information processing system 200. For example, the information processing system 200 can be applied to a case where each subspace forming a public space where people gather is set as a target area and whether or not dangerous congestion has occurred in each subspace.

[0050] Specifically, when it is determined that dangerous congestion has occurred in any of the partial spaces, the information processing system 200 may issue an alarm to a manager who manages safety in the public space, or to a worker who actually manages the flow of people in any of the partial spaces. This allows the information processing system 200 to make it easier to detect dangerous congestion in the public space early in real time, to make it easier to prevent crowd accidents, and to make it easier to ensure safety.

[0051] Furthermore, the information processing system 200 may specifically determine whether or not dangerous congestion occurred in each subspace for each past time period, and analyze which subspaces are likely to experience dangerous congestion during which time periods. This allows the information processing system 200 to easily grasp which time periods are likely to experience dangerous congestion in public spaces, and facilitates measures to prevent crowd accidents.

[0052] (Example of hardware configuration of information processing device 100) Next, an example of the hardware configuration of the information processing device 100 will be described with reference to FIG.

[0053] Fig. 3 is a block diagram showing an example of the hardware configuration of the information processing device 100. In Fig. 3, the information processing device 100 has a CPU (Central Processing Unit) 301, a memory 302, and a network I / F (Interface) 303. The information processing device 100 also has a recording medium I / F 304, a recording medium 305, a display 306, and an input device 307. The components are connected to each other via a bus 300.

[0054] Here, CPU 301 is responsible for overall control of information processing device 100. Memory 302 includes, for example, a read-only memory (ROM), a random access memory (RAM), and a flash ROM. Specifically, for example, the flash ROM or ROM stores various programs, and RAM is used as a work area for CPU 301. The programs stored in memory 302 are loaded into CPU 301, causing CPU 301 to execute coded processes.

[0055] The network I / F 303 is connected to the network 210 via a communication line, and is connected to other computers via the network 210. The network I / F 303 manages the internal interface with the network 210 and controls the input and output of data from other computers. The network I / F 303 is, for example, a modem or a LAN adapter.

[0056] The recording medium I / F 304 controls reading and writing of data from and to the recording medium 305 under the control of the CPU 301. The recording medium I / F 304 is, for example, a disk drive, a solid state drive (SSD), or a universal serial bus (USB) port. The recording medium 305 is a non-volatile memory that stores data written under the control of the recording medium I / F 304. The recording medium 305 is, for example, a disk, a semiconductor memory, or a USB memory. The recording medium 305 may be detachable from the information processing device 100.

[0057] The display 306 displays data such as a cursor, an icon, a toolbox, a document, an image, or function information. The display 306 is, for example, a CRT (Cathode Ray Tube), a liquid crystal display, or an organic EL (Electroluminescence) display. The input device 307 has keys for inputting characters, numbers, various instructions, etc., and inputs data. The input device 307 is, for example, a keyboard or a mouse. The input device 307 may also be, for example, a touch panel input pad or a numeric keypad.

[0058] The information processing device 100 may have, in addition to the above-described components, for example, a camera. The information processing device 100 may have, in addition to the above-described components, for example, a printer, a scanner, a microphone, or a speaker. The information processing device 100 may have, for example, a plurality of recording medium I / Fs 304 and recording media 305. The information processing device 100 may not have, for example, a display 306 or an input device 307. The information processing device 100 may not have, for example, a recording medium I / F 304 or a recording medium 305.

[0059] (Example of hardware configuration of imaging device 201) Next, an example of the hardware configuration of the imaging device 201 will be described with reference to FIG.

[0060] Fig. 4 is a block diagram showing an example of the hardware configuration of the image capture device 201. In Fig. 4, the image capture device 201 has a CPU 401, a memory 402, a network I / F 403, a recording medium I / F 404, a recording medium 405, and a camera 406. Furthermore, each component is connected to each other via a bus 400.

[0061] Here, the CPU 401 is responsible for overall control of the imaging device 201. The memory 402 includes, for example, a ROM, a RAM, and a flash ROM. Specifically, for example, the flash ROM and the ROM store various programs, and the RAM is used as a work area for the CPU 401. The programs stored in the memory 402 are loaded into the CPU 401, causing the CPU 401 to execute coded processes.

[0062] The network I / F 403 is connected to the network 210 via a communication line, and is connected to other computers via the network 210. The network I / F 403 manages an internal interface with the network 210 and controls input and output of data from other computers. The network I / F 403 is, for example, a modem or a LAN adapter.

[0063] The recording medium I / F 404 controls reading / writing of data from / to the recording medium 405 under the control of the CPU 401. The recording medium I / F 404 is, for example, a disk drive, an SSD, a USB port, etc. The recording medium 405 is a non-volatile memory that stores data written under the control of the recording medium I / F 404. The recording medium 405 is, for example, a disk, a semiconductor memory, a USB memory, etc. The recording medium 405 may be detachable from the imaging device 201.

[0064] The camera 406 has multiple image sensors and generates an image of a specific area using the multiple image sensors. For example, if a person is present in the specific area, the camera 406 generates an image that captures the person. The camera 406 is, for example, a fixed camera. The camera 406 may be, for example, movable. The camera 406 is, for example, a surveillance camera.

[0065] In addition to the components described above, the image capturing device 201 may also include, for example, a keyboard, a mouse, a display, a printer, a scanner, a microphone, a speaker, etc. The image capturing device 201 may also include a plurality of recording medium I / Fs 404 and recording media 405. The image capturing device 201 may also not include a recording medium I / F 404 or a recording medium 405.

[0066] (Example of hardware configuration of client device 202) A specific example of the hardware configuration of the client device 202 is similar to the example of the hardware configuration of the information processing device 100 shown in FIG. 3, and therefore a description thereof will be omitted.

[0067] (Example of functional configuration of information processing device 100) Next, an example of the functional configuration of the information processing device 100 will be described with reference to FIG.

[0068] 5 is a block diagram showing an example of the functional configuration of the information processing device 100. The information processing device 100 includes a storage unit 500, an acquisition unit 501, a generation unit 502, a learning unit 503, an analysis unit 504, and an output unit 505.

[0069] The storage unit 500 is realized by, for example, a storage area such as the memory 302 or the recording medium 305 shown in Fig. 3. In the following, a case where the storage unit 500 is included in the information processing device 100 will be described, but this is not limiting. For example, the storage unit 500 may be included in a device different from the information processing device 100, and the stored contents of the storage unit 500 may be accessible from the information processing device 100.

[0070] The acquiring unit 501 to the output unit 505 function as an example of a control unit. Specifically, the acquiring unit 501 to the output unit 505 realize their functions by causing the CPU 301 to execute a program stored in a storage area such as the memory 302 or the recording medium 305 shown in Fig. 3, or by using the network I / F 303. The processing results of each functional unit are stored in a storage area such as the memory 302 or the recording medium 305 shown in Fig. 3, for example.

[0071] The storage unit 500 stores various information that is referenced or updated in the processing of each functional unit. The storage unit 500 stores, for example, multiple pieces of position data indicating the positions of each moving object in a target area at different times. The moving object is, for example, a person. The moving object may be, for example, a ship, a vehicle, an airplane, or a drone. The position data indicates the position of each moving object in a target area, for example, without identifying the individual. In the following explanation, a case will be described in which the moving object is a person. The position data indicates the position of each person in a target area, for example, without identifying the individual. There may be, for example, multiple target areas.

[0072] Specifically, the storage unit 500 stores a plurality of pieces of position data indicating the positions of people in a target area at different times within one or more periods. The position data is acquired by, for example, the acquisition unit 501.

[0073] The storage unit 500 may store, for example, a correct answer as a result of analyzing the risk due to the congestion situation in the target area for each of one or more periods. The correct answer is, for example, whether or not there is a risk. Specifically, the storage unit 500 may store, as the correct answer, information indicating whether or not there is a risk due to the congestion situation in the target area for each of one or more periods. The correct answer is acquired, for example, by the acquisition unit 501.

[0074] Specifically, the storage unit 500 stores a plurality of position data indicating the positions of people in a target area at different times within a predetermined period. The predetermined period is a period during which it is desired to analyze the risk of congestion in the target area. The position data is acquired by, for example, the acquisition unit 501.

[0075] The storage unit 500 stores, for example, a model that realizes a function of outputting a result of analyzing the risk due to the congestion situation in a target area according to input feature data that represents spatial characteristics in the target area. The result of the risk analysis is, for example, the presence or absence of risk. The model is, for example, a neural network. The model may be, for example, a decision tree. The model may be, for example, a mathematical formula. Specifically, the model may be, for example, a mathematical formula that enables classification of a cluster of feature data corresponding to the presence of risk and a cluster of feature data corresponding to the absence of risk. The model is trained, for example, by the learning unit 503.

[0076] The acquisition unit 501 acquires various types of information used in processing by each functional unit. The acquisition unit 501 stores the acquired various types of information in the storage unit 500 or outputs it to each functional unit. The acquisition unit 501 may also output the various types of information stored in the storage unit 500 to each functional unit. The acquisition unit 501 acquires various types of information based on, for example, a user's operation input. The acquisition unit 501 may receive various types of information from, for example, a device different from the information processing device 100.

[0077] The acquisition unit 501 acquires, for example, a plurality of pieces of position data indicating the positions of each person in a target area at different times within one or more periods. Specifically, the acquisition unit 501 acquires the plurality of pieces of position data by receiving the plurality of pieces of position data from another computer. The other computer is, for example, the imaging device 201. Specifically, the acquisition unit 501 may acquire the plurality of pieces of position data by accepting input of the plurality of pieces of position data.

[0078] Specifically, the acquiring unit 501 may acquire images of the target area captured at different times within each of one or more periods. More specifically, the acquiring unit 501 acquires the images by receiving the images from another computer. Specifically, the acquiring unit 501 may analyze the acquired images to acquire a plurality of position data indicating the positions of each person in the target area at different times within each of the one or more periods.

[0079] The acquiring unit 501 acquires a correct answer as a result of analyzing the risk due to the congestion situation in the target area for each of one or more periods, for example. The correct answer is, for example, whether there is a risk. Specifically, the acquiring unit 501 acquires the information by, for example, receiving information indicating whether there is a risk due to the congestion situation in the target area for each of one or more periods from another computer. The other computer is, for example, the client device 202. Specifically, the acquiring unit 501 may acquire the information by, for example, accepting input of information indicating whether there is a risk due to the congestion situation in the target area for each of one or more periods.

[0080] The acquiring unit 501 acquires, for example, a plurality of pieces of position data indicating the positions of people in a target area at different times within a predetermined period. Specifically, the acquiring unit 501 acquires the plurality of pieces of position data by receiving the plurality of pieces of position data from another computer. The other computer is, for example, the imaging device 201. Specifically, the acquiring unit 501 may acquire the plurality of pieces of position data by accepting input of the plurality of pieces of position data.

[0081] Specifically, the acquiring unit 501 may acquire images of the target area captured at different times within a predetermined period. More specifically, the acquiring unit 501 acquires the images by receiving the images from another computer. Specifically, the acquiring unit 501 may analyze the acquired images to acquire multiple pieces of position data indicating the positions of people in the target area at different times within the predetermined period.

[0082] The acquisition unit 501 may receive a start trigger that starts processing of any of the functional units. The start trigger may be, for example, a predetermined operation input by a user. The start trigger may be, for example, reception of predetermined information from another computer. The start trigger may be, for example, output of predetermined information by any of the functional units.

[0083] Specifically, the acquiring unit 501 accepts the acquisition of multiple pieces of position data relating to each of one or more periods as a start trigger for starting the processing of the generating unit 502 and the learning unit 503. Specifically, the acquiring unit 501 accepts the acquisition of multiple pieces of position data indicating the positions of each person in the target area at different times within a predetermined period as a start trigger for starting the processing of the generating unit 502 and the analyzing unit 504.

[0084] The generating unit 502 generates feature amount data representing spatial features of the target area through topological data analysis based on the plurality of position data acquired by the acquiring unit 501. The feature amount data is, for example, data based on PI. For example, the generating unit 502 generates first feature amount data for each period based on the plurality of position data for each of one or more periods.

[0085] Specifically, the generation unit 502 generates a diagram based on the acquired multiple pieces of position data for each period. The diagram represents the timing at which one or more different types of shapes appear and disappear depending on the change in resolution. Specifically, the diagram is a PD. Each of the one or more shapes is a shape that can be formed by a combination of people's positions on the target area. Specifically, the one or more shapes include at least one of a connecting shape, a ring shape, and a hollow shape that can be formed by a combination of people's positions on the target area.

[0086] Specifically, the generating unit 502 generates first feature amount data representing spatial features in the target area corresponding to the generated diagram for each period. More specifically, the generating unit 502 generates a PI corresponding to the generated PD and generates first feature amount data corresponding to the generated PI. This enables the generating unit 502 to train a model that realizes a function of outputting an analysis result of the risk due to the congestion situation in the target area, taking into account the spatial features in the target area, and can improve the accuracy of the trained model.

[0087] The learning unit 503 learns a model based on the first feature amount data generated by the generation unit 502. For example, for each period, the learning unit 503 identifies a combination of the first feature amount data and a correct answer as a result of analyzing the risk due to the congestion situation in the target area during that period. The learning unit 503 learns a model, for example, by supervised learning based on the identified combination. For example, the learning unit 503 may also learn a model by unsupervised learning based on the first feature amount data. This allows the learning unit 503 to learn a model that can accurately analyze the risk due to the congestion situation in the target area, taking into account the spatial features of the target area.

[0088] The generating unit 502 generates second feature amount data for a predetermined period based on, for example, a plurality of position data for the predetermined period. Specifically, the generating unit 502 generates a diagram based on the plurality of position data acquired for the predetermined period. Specifically, the generating unit 502 generates second feature amount data representing spatial features in a target area corresponding to the generated diagram for the predetermined period. More specifically, the generating unit 502 generates a PI corresponding to the generated PD and generates second feature amount data corresponding to the generated PI. In this way, the generating unit 502 can make a model usable taking into account spatial features in the target area.

[0089] The analysis unit 504 acquires the result of analyzing the risk of congestion in the target area for a predetermined period of time based on the second feature amount data generated by the generation unit 502, using the model trained by the training unit 503. The analysis unit 504 acquires the result of analyzing the risk of congestion in the target area for a predetermined period of time, for example, by inputting the second feature amount data into the model. This enables the analysis unit 504 to accurately analyze the risk of congestion in the target area for a predetermined period of time using the model.

[0090] The output unit 505 outputs the processing result of at least one of the functional units. The output format is, for example, display on a display, printout to a printer, transmission to an external device via the network I / F 303, or storage in a storage area such as the memory 302 or the recording medium 305. This enables the output unit 505 to notify the user of the processing result of at least one of the functional units, and can support the management and operation of the information processing device 100, for example, updating setting values ​​of the information processing device 100, thereby improving the convenience of the information processing device 100.

[0091] The output unit 505 transmits, for example, the result of the analysis of the risk due to the congestion situation in the target area for a predetermined period acquired by the analysis unit 504 to another computer. The other computer is, for example, the client device 202. The output unit 505 may output, for example, the result of the analysis of the risk due to the congestion situation in the target area for a predetermined period acquired by the analysis unit 504 so that it can be referenced by a user. In this way, the output unit 505 can make the risk due to the congestion situation in the target area for a predetermined period available for external reference.

[0092] The output unit 505 may output, for example, the model learned by the learning unit 503. Specifically, the output unit 505 transmits the model learned by the learning unit 503 to another computer. The other computer is, for example, the client device 202. In this way, the output unit 505 can make available to the outside a model that can accurately analyze risks due to the congestion situation in the target area.

[0093] (An example of the operation of the information processing device 100) Next, an example of the operation of the information processing device 100 will be described with reference to FIGS.

[0094] 6 to 12 are explanatory diagrams showing an example of the operation of the information processing device 100. In FIGS. 6 to 12, the information processing device 100 determines whether or not dangerous congestion has occurred in a target area. The target area is a two-dimensional plane onto which a three-dimensional space is projected, excluding the height component. The two-dimensional plane includes an X axis and a Y axis. A position is a combination of an X coordinate and a Y coordinate.

[0095] Here, for example, when dangerous congestion occurs, a unique state may emerge regarding the spatial characteristics of the target area and the persistence of the spatial characteristics. For example, when an arch-shaped crowd forms at a narrow section of the target area and a dangerous congestion occurs, a unique state may emerge in which the arch-shaped crowd is maintained over time and the area of ​​the congestion expands from the arch-shaped crowd. For this reason, the spatial characteristics of the target area and the state regarding the persistence of the spatial characteristics are considered to be guidelines for determining whether dangerous congestion has occurred. However, it may be difficult to define spatial characteristics in advance.

[0096] Therefore, the information processing device 100 uses topological data analysis to learn the state that appears in the spatial features of the target area and the continuity of the spatial features when dangerous congestion occurs, based on point cloud data indicating the positions of the crowd, without tracking each person in the target area. For example, the information processing device 100 uses topological data analysis to generate feature amount data that reflects the spatial features of the target area and the continuity of the spatial features, based on the point cloud data indicating the positions of the crowd. For example, the information processing device 100 uses the generated feature amount data to learn a model that realizes a function of determining whether dangerous congestion has occurred.

[0097] The model may be, for example, a neural network. The model may be, for example, a decision tree. The model may be, for example, a mathematical formula. Specifically, the model may be, for example, a mathematical formula that enables classification into a cluster of feature data corresponding to a case where dangerous congestion occurs and a cluster of feature data corresponding to a case where dangerous congestion does not occur. The model is trained, for example, by unsupervised learning. The model may also be trained, for example, by supervised learning.

[0098] An example in which the information processing device 100 determines whether dangerous congestion has occurred in a target area will be described below with reference to Fig. 6 to Fig. 12. First, the description will shift to Fig. 6.

[0099] In Fig. 6, the information processing device 100 acquires point cloud data 600 corresponding to each time point within a first period. The point cloud data 600 indicates the position of a crowd in a target area at a certain time point. The example of Fig. 6 shows the point cloud data 600 at time point t = 0. The information processing device 100 applies kernel density estimation to the point cloud data 600 corresponding to each time point to generate 2D pixel data 610 corresponding to each time point. Next, we move on to the description of Fig. 7.

[0100] 7, the information processing device 100 generates 3D voxel data 700 corresponding to a first time period by stacking 2D pixel data 610 corresponding to each time point in the time direction. This allows the information processing device 100 to prepare the 3D voxel data 700 that serves as the basis for analyzing the spatial characteristics of a target region and the state of continuity of the spatial characteristics.

[0101] Next, the information processing device 100 generates a PD that represents the spatial characteristics of the target region and the state of continuity of the spatial characteristics, corresponding to the 3D voxel data 700. The PD represents the timing at which each of one or more different types of shapes appears and disappears depending on the change in resolution. Each of the one or more shapes is a shape that can be formed by a combination of people's positions. For example, each of the one or more shapes is a shape that can be formed by a combination of spheres centered on the person's position. Specifically, the one or more shapes include at least one of a connected shape, a ring shape, and a hollow shape. Now, moving on to the description of Figures 8 and 9, an example of a PD will be described.

[0102] In the example of FIG. 8, it is assumed that a point cloud 800 exists. The information processing device 100 identifies the timing at which ring shapes that can be formed by a combination of spheres appear and disappear, depending on the change in resolution r for each point of the point cloud 800. The resolution r indicates the size at which each point is recognized. The resolution r is the radius of a sphere centered at each point. In the example of FIG. 8, as indicated by reference numerals 811 to 813, a ring α1 appears at r=2. As indicated by reference numerals 811 to 813, a ring α2 appears at r=2 and disappears at r=2.5.

[0103] As shown in Fig. 9, the information processing device 100 determines the timing at which the ring shape appears and disappears in accordance with changes in resolution r, and generates PD900 representing each timing. The horizontal axis of PD900 indicates the timing at which the ring appears. The vertical axis of PD900 indicates the timing at which the ring disappears. Now, we move on to the explanation of Fig. 10.

[0104] 10, the information processing device 100 specifically generates a PD1000 corresponding to the 3D voxel data 700. The PD1000 indicates the timing at which a connection shape appears and disappears, the timing at which a ring shape appears and disappears, and the timing at which a hollow shape appears and disappears.

[0105] H0 corresponds to a connection shape. H1 corresponds to a ring shape. H2 corresponds to a cavity shape. Birth corresponds to emergence. Death corresponds to disappearance. Here, for the processing procedure for generating PD1000 based on 3D voxel data 700, for example, reference can be made to Reference 2 below.

[0106] Reference 2: Kaji, Shizuo, Takeki Sudo, and Kazushi Ahara. “Cubical ripser: Software for computing persistent homology of image and volume data.” arXiv preprint arXiv:2005.12692 (2020).

[0107] Next, we will move on to the description of Fig. 11 and Fig. 12. In Fig. 11 and Fig. 12, the information processing device 100 generates a PI that indicates a frequency distribution corresponding to the timing at which each of one or more shapes of different types appears and disappears in the generated PD 1000. Here, an example of the PI will be described with reference to Fig. 11.

[0108] As shown in Fig. 11, assume that PD1100 exists. 1110 indicates the frequency corresponding to the timing at which a ring shape appears and disappears, corresponding to PD1100. PI1120 indicates the frequency distribution corresponding to 1110. In reality, the information processing device 100 generates PI1120 from PD1100 using a distribution function, without using the histogram shown in 1110. Now, we move on to the description of Fig. 12.

[0109] 12, the information processing device 100 specifically generates PI1210 indicating a frequency distribution corresponding to the timing at which a connected shape appears and disappears in the generated PD 1000. Specifically, the information processing device 100 generates PI1220 indicating a frequency distribution corresponding to the timing at which a ring shape appears and disappears in the generated PD 1000. Specifically, the information processing device 100 generates PI1230 indicating a frequency distribution corresponding to the timing at which a hollow shape appears and disappears in the generated PD 1000.

[0110] The information processing device 100 generates a feature vector for the first period based on the generated PIs 1210, 1220, and 1230. Similarly, the information processing device 100 generates a feature vector for each of one or more periods other than the first period. Based on the generated feature vectors, the information processing device 100 learns, by unsupervised learning or supervised learning, a model that realizes a function of determining whether dangerous congestion has occurred in a target area according to the feature vector.

[0111] The information processing device 100 uses the trained model to determine whether dangerous congestion has occurred in a target area during a predetermined period. The information processing device 100 generates a feature vector for the predetermined period based on point cloud data corresponding to each time point within the predetermined period. The information processing device 100 determines whether dangerous congestion has occurred in a target area during a predetermined period by, for example, inputting the generated feature vector into the trained model.

[0112] The information processing device 100 outputs the result of determining whether dangerous congestion has occurred in a target area during a predetermined period. For example, the information processing device 100 transmits the result of determining whether dangerous congestion has occurred in a target area during a predetermined period to the client device 202. For example, the information processing device 100 may output the result of determining whether dangerous congestion has occurred in a target area during a predetermined period so that the user can refer to it.

[0113] This allows the information processing device 100 to learn a model capable of appropriately determining whether dangerous congestion has occurred. For example, the information processing device 100 can learn a model capable of appropriately determining whether dangerous congestion has occurred by distinguishing between dangerous congestion and safe congestion in consideration of spatial characteristics when dangerous congestion occurs using the PD1000.

[0114] Therefore, the information processing device 100 can appropriately determine whether or not dangerous congestion has occurred, specifically, regardless of the density of people in the target area. Specifically, the information processing device 100 can appropriately determine whether or not dangerous congestion has occurred, even if the target area is a relatively large space or has a relatively complex shape, and the density of people in the target area is not uniform.

[0115] Specifically, the information processing device 100 can appropriately determine whether dangerous congestion has occurred in a public space where people gather, which is actually a relatively large space or has a relatively complex shape. Furthermore, specifically, even if the density of people in the target area is relatively high, the information processing device 100 can avoid erroneously determining that dangerous congestion has occurred if the flow of people is regular and only safe congestion has occurred.

[0116] Specifically, the information processing device 100 can appropriately determine whether dangerous congestion has occurred without identifying individuals or tracking the movement of each person in a target area. Specifically, the information processing device 100 can determine whether dangerous congestion has occurred without tracking the movement of people in a target area, thereby suppressing an increase in processing time or processing load.

[0117] Therefore, the information processing device 100 can make it easier to prevent crowd accidents. For example, the information processing device 100 can make it possible to appropriately determine whether dangerous congestion that could lead to a crowd accident has occurred in order to prevent a crowd accident. The information processing device 100 can also make it possible for a manager who manages safety in a target area or a worker who manages the flow of people in the target area to quickly grasp that dangerous congestion that could lead to a crowd accident has occurred. The information processing device 100 can make it easier for a manager or a worker to quickly take appropriate measures against dangerous congestion that has occurred.

[0118] Furthermore, when safe congestion is occurring but dangerous congestion is not, the information processing device 100 can avoid notifying a manager or worker, etc., of an alarm indicating the occurrence of dangerous congestion. Therefore, the information processing device 100 can avoid issuing excessively frequent alarms to the manager or worker, etc. The information processing device 100 can suppress an increase in the physical or psychological workload imposed on the manager or worker, etc. Furthermore, the information processing device 100 can prevent the manager or worker, etc., from falling into a psychological state that makes them disregard the alarm.

[0119] (Specific example of operation of information processing device 100) Next, a specific example of the operation of the information processing device 100 will be described with reference to FIGS.

[0120] 13 and 14 are explanatory diagrams showing a specific example of the operation of the information processing device 100. In Fig. 13, the information processing device 100 acquires a point cloud dataset 1300 by receiving multiple point cloud data groups relating to different time periods from the imaging device 201. The multiple point cloud data groups include, for example, point cloud data group 1 and point cloud data group 2. The multiple point cloud data groups include, for example, target point cloud data for determining whether dangerous congestion has occurred during a predetermined time period. The information processing device 100 acquires, for example, information indicating whether dangerous congestion has occurred in a target area during each of the different time periods.

[0121] The information processing device 100 generates 3D voxel data for each period of time for a point cloud data group related to the period. The information processing device 100 generates a PD corresponding to the generated 3D voxel data for each period of time. The information processing device 100 generates a PI corresponding to the generated PD for each period of time. The information processing device 100 generates a feature vector 1310 based on the generated PI for each period of time.

[0122] The information processing device 100 generates a two-dimensional feature vector 1320 by reducing the dimension of the generated feature vector 1310 for each period using PCA (Principal Component Analysis). Based on the generated feature vector 1320 for each period, the information processing device 100 classifies a plurality of point cloud data groups into a congested cluster A and a non-congested cluster B using 2-means. Now, we move on to the description of FIG. 14.

[0123] The open and closed circles in graph 1400 in Fig. 14 represent point cloud data groups projected onto a two-dimensional plane by PCA. The open circles correspond to point cloud data groups relating to periods when dangerous congestion does not occur. The closed circles correspond to point cloud data groups relating to periods when dangerous congestion occurs. In this way, the information processing device 100 can appropriately classify multiple point cloud data groups into a crowded cluster A and a non-crowded cluster B by taking spatial characteristics into account via PD.

[0124] Here, when the information processing device 100 classifies the target point cloud data group into congestion cluster A, it determines that dangerous congestion has occurred in the target area during a predetermined period. When the information processing device 100 classifies the target point cloud data group into non-congestion cluster B, it determines that dangerous congestion has not occurred in the target area during a predetermined period. When the information processing device 100 determines that dangerous congestion has occurred in the target area during a predetermined period, it outputs an alert indicating that dangerous congestion has occurred in the target area. The information processing device 100 transmits the alert to the client device 202, for example.

[0125] This allows the information processing device 100 to learn a model that can appropriately determine whether dangerous congestion has occurred. For example, the information processing device 100 can learn a model that can appropriately determine whether dangerous congestion has occurred by distinguishing between dangerous congestion and safe congestion, taking into account spatial characteristics when dangerous congestion occurs using PD.

[0126] Therefore, the information processing device 100 can appropriately determine whether or not dangerous congestion has occurred, specifically, regardless of the density of people in the target area. Specifically, the information processing device 100 can appropriately determine whether or not dangerous congestion has occurred, even if the target area is a relatively large space or has a relatively complex shape, and the density of people in the target area is not uniform.

[0127] Specifically, the information processing device 100 can appropriately determine whether dangerous congestion has occurred in a public space where people gather, which is actually a relatively large space or has a relatively complex shape. Furthermore, specifically, even if the density of people in the target area is relatively high, the information processing device 100 can avoid erroneously determining that dangerous congestion has occurred if the flow of people is regular and only safe congestion has occurred.

[0128] Specifically, the information processing device 100 can appropriately determine whether dangerous congestion has occurred without identifying individuals or tracking the movement of each person in a target area. Specifically, the information processing device 100 can determine whether dangerous congestion has occurred without tracking the movement of people in a target area, thereby suppressing an increase in processing time or processing load.

[0129] Therefore, the information processing device 100 can make it easier to prevent crowd accidents. For example, the information processing device 100 can make it possible to appropriately determine whether dangerous congestion that could lead to a crowd accident has occurred in order to prevent a crowd accident. The information processing device 100 can also make it possible for a manager who manages safety in a target area or a worker who manages the flow of people in the target area to quickly grasp that dangerous congestion that could lead to a crowd accident has occurred. The information processing device 100 can make it easier for a manager or a worker to quickly take appropriate measures against dangerous congestion that has occurred.

[0130] Furthermore, when safe congestion is occurring but dangerous congestion is not, the information processing device 100 can avoid notifying a manager or worker, etc., of an alarm indicating the occurrence of dangerous congestion. Therefore, the information processing device 100 can avoid issuing excessively frequent alarms to the manager or worker, etc. The information processing device 100 can suppress an increase in the physical or psychological workload imposed on the manager or worker, etc. Furthermore, the information processing device 100 can prevent the manager or worker, etc., from falling into a psychological state that makes them disregard the alarm.

[0131] (Example of output screen) Next, an example of an output screen will be described with reference to FIG.

[0132] Fig. 15 is an explanatory diagram showing an example of an output screen. In Fig. 15, the information processing device 100 determines whether dangerous congestion has occurred in each of the individual areas 1501 to 1506 of the overall area 1500, based on a group of point cloud data received from the imaging devices 201 provided in each of the individual areas 1501 to 1506.

[0133] Here, it is assumed that the information processing device 100 determines that dangerous congestion has occurred in the individual area 1504. The information processing device 100 generates an output screen 1510 including a notification indicating that dangerous congestion has occurred in the individual area 1504, so that the individual area 1504 can be identified on the map of the entire area 1500. The information processing device 100 transmits the generated output screen to the client device 202.

[0134] This makes it easier for the information processing device 100 to prevent crowd accidents. For example, the information processing device 100 enables an administrator who manages safety in the entire area 1500 or an operator who physically manages the flow of people in the individual area 1504 to grasp an individual area 1504 in which dangerous congestion that could lead to a crowd accident has occurred. The information processing device 100 makes it easier for, for example, an administrator or an operator to take appropriate measures early on in response to the dangerous congestion that has occurred.

[0135] Here, the case where the information processing device 100 generates the output screen 1510 has been described, but the present invention is not limited to this. For example, the information processing device 100 may transmit the result of its determination that dangerous congestion has occurred in the individual area 1504 to the client device 202. In this case, when the client device 202 receives the result of its determination that dangerous congestion has occurred in the individual area 1504 from the information processing device 100, it generates the output screen 1510.

[0136] (Another example of the output screen) Next, another example of the output screen will be described. The information processing device 100 may determine whether dangerous congestion occurred in a target area during each of a plurality of individual periods within a past overall period. The overall period is, for example, one day. If the overall period is one day, the individual periods are, for example, each time slot per hour. The overall period may be, for example, one week. If the overall period is one week, the individual periods are, for example, each day of the week.

[0137] Specifically, the information processing device 100 may determine whether dangerous congestion has occurred in a target area for each hourly time slot in a day. The information processing device 100 may then generate an output screen that associates the results of determining whether dangerous congestion has occurred in a target area for each time slot on a time axis for one day, and transmit the output screen to the client device 202.

[0138] Specifically, the information processing device 100 may determine whether dangerous congestion has occurred in a target area for each hour of the day for each of a plurality of days. Then, specifically, the information processing device 100 may calculate the frequency with which dangerous congestion occurs in the target area for each time period of the day based on the result of determining whether dangerous congestion has occurred in the target area. Specifically, the information processing device 100 may generate an output screen in which the frequency with which dangerous congestion occurs in the target area for each time period is associated on a time axis for one day, and transmit the output screen to the client device 202.

[0139] As a result, the information processing device 100 enables a manager who manages safety in a target area or a worker who manages the flow of people in the target area to understand during what time of day dangerous congestion is likely to occur in the target area. Therefore, the information processing device 100 makes it easier for a manager or a worker to take appropriate measures early on regarding time periods when dangerous congestion is likely to occur.

[0140] (Overall processing procedure) Next, an example of an overall processing procedure executed by the information processing device 100 will be described with reference to Fig. 16. The overall processing is realized by, for example, the CPU 301, storage areas such as the memory 302 and the recording medium 305, and the network I / F 303 shown in Fig. 3.

[0141] Fig. 16 is a flowchart showing an example of an overall processing procedure. In Fig. 16, the information processing device 100 acquires a point cloud dataset that compiles point cloud data indicating the positions of each person in a crowd present in a target area at each time point within each period including a predetermined period (step S1601).

[0142] The information processing device 100 applies kernel density estimation to each point cloud data of the acquired point cloud dataset for each period and stacks the data in the time direction to generate 3D voxel data (step S1602).The information processing device 100 generates a PD corresponding to the generated 3D voxel data for each period (step S1603).

[0143] The information processing device 100 converts the generated PD for each period into PI to generate a feature vector (step S1604). The information processing device 100 learns a model that realizes a function of determining whether or not there is a risk due to the congestion situation in the target area based on the generated feature vector for each period other than the predetermined period (step S1605).

[0144] The information processing device 100 uses the learned model to determine whether or not there is a risk due to the congestion situation in the target area for a predetermined period of time, based on the generated feature vector (step S1606). The information processing device 100 then ends the overall processing.

[0145] As described above, the information processing device 100 can acquire multiple pieces of position data indicating the positions of each moving object in a target area at different times. The information processing device 100 can generate a diagram showing the timing at which one or more different types of shapes appear and disappear depending on changes in resolution, based on the acquired multiple pieces of position data through topological data analysis. The information processing device 100 can generate first feature data representing spatial characteristics of the target area corresponding to the diagram. The information processing device 100 can train a model that outputs an analysis result of the risk due to congestion in the target area based on the generated first feature data and the input feature data representing the spatial characteristics of the target area. This allows the information processing device 100 to appropriately analyze the risk due to congestion in the target area.

[0146] According to the information processing device 100, people can be used as moving objects, which makes it easier for the information processing device 100 to prevent crowd accidents.

[0147] The information processing device 100 can generate diagrams for one or more shapes that can be formed by a combination of the positions of moving objects in a target area, including at least any of a connected shape, a ring shape, and a hollow shape. This allows the information processing device 100 to learn a model that can accurately analyze the risk of congestion in the target area by appropriately considering the spatial characteristics of the target area.

[0148] The information processing device 100 can acquire, for each of a plurality of time periods, a plurality of position data indicating the position of each moving object in a target area at different times within the time period. The information processing device 100 can generate, for each time period, a diagram showing the timing at which each shape appears and disappears depending on changes in resolution, based on the acquired plurality of position data through topological data analysis. The information processing device 100 can generate, for each time period, first feature amount data representing spatial characteristics in the target area corresponding to the diagram. This allows the information processing device 100 to prepare a plurality of first feature amount data to be used when training a model, making it easier to train the model with high accuracy.

[0149] According to the information processing device 100, a model can be learned based on a combination of the generated first feature amount data and the presence or absence of risk due to the congestion situation in the target area for each period. This enables the information processing device 100 to easily learn a model with high accuracy through supervised learning.

[0150] The information processing device 100 can acquire multiple position data indicating the positions of each moving object in a target area at different times within a predetermined period. The information processing device 100 can generate a diagram showing the timing at which each shape appears and disappears depending on changes in resolution based on the multiple position data through topological data analysis. The information processing device 100 can generate second feature data representing spatial characteristics of the target area corresponding to the diagram. The information processing device 100 can output results of an analysis of the risk of congestion in the target area over a predetermined period based on the generated second feature data using a trained model. This allows the information processing device 100 to accurately analyze the risk of congestion in the target area over a predetermined period using the trained model.

[0151] The information processing device 100 can output the learned model, thereby making the model capable of accurately analyzing risks due to the congestion state in a target area available to the outside.

[0152] The information processing method described in this embodiment can be realized by executing a prepared program on a computer such as a PC or a workstation. The information processing program described in this embodiment is recorded on a computer-readable recording medium and executed by being read from the recording medium by the computer. The recording medium may be a hard disk, a flexible disk, a CD (Compact Disc)-ROM, an MO (Magneto Optical disc), a DVD (Digital Versatile Disc), or the like. The information processing program described in this embodiment may also be distributed via a network such as the Internet.

[0153] The following additional notes are provided regarding the above-described embodiment.

[0154] (Appendix 1) Acquire a plurality of position data indicating the positions of each moving object in the target area at different times; generating first feature amount data representing spatial features of the area, the first feature amount data corresponding to a diagram representing timings at which one or more shapes of different types that can be formed by combinations of positions of moving objects on the area appear and disappear according to changes in resolution, based on the plurality of position data acquired by topological data analysis; learning a model that outputs a result of analyzing a risk due to a congestion situation in the area in accordance with input feature amount data that represents spatial characteristics of the area, based on the generated first feature amount data; An information processing program that causes a computer to execute a process.

[0155] (Supplementary Note 2) The information processing program according to Supplementary Note 1, wherein the moving body is a person.

[0156] (Appendix 3) The information processing program described in Appendix 2, characterized in that the one or more shapes include at least one of a connected shape, a ring shape, or a hollow shape that can be formed by a combination of positions of moving objects on the area.

[0157] (Appendix 4) The acquisition process is acquiring, for each of a plurality of time periods, a plurality of position data indicating the position of each of the moving objects in the area at different times within the time period; The generating process includes: The information processing program according to claim 3, further comprising: generating, for each of the periods based on the acquired position data through topological data analysis, first feature data representing spatial characteristics of the region, corresponding to a diagram showing the timing at which each of the shapes appears and disappears according to a change in resolution.

[0158] (Appendix 5) The learning process is The information processing program according to any one of appendices 1 to 4, characterized in that the model is trained based on a combination of the generated first feature data and the presence or absence of risk due to the congestion situation in the area during each of the periods.

[0159] (Supplementary Note 6) By topological data analysis, based on a plurality of position data indicating the positions of each moving object on the area at different times within a predetermined period, second feature amount data representing spatial features on the area is generated, the second feature amount data corresponding to a diagram showing the timing at which each shape appears and disappears depending on a change in resolution; outputting a result of analyzing the risk due to the congestion situation in the area during the predetermined period based on the generated second feature amount data using the trained model; 5. The information processing program according to any one of claims 1 to 4, which causes the computer to execute the process.

[0160] (Appendix 7) Output the learned model. 5. The information processing program according to any one of claims 1 to 4, which causes the computer to execute the process.

[0161] (Appendix 8) Acquire a plurality of position data indicating the positions of each moving object in the target area at different times; generating first feature amount data representing spatial features of the area, the first feature amount data corresponding to a diagram representing timings at which one or more shapes of different types that can be formed by combinations of positions of moving objects on the area appear and disappear according to changes in resolution, based on the plurality of position data acquired by topological data analysis; learning a model that outputs a result of analyzing a risk due to a congestion situation in the area in accordance with input feature amount data that represents spatial characteristics of the area, based on the generated first feature amount data; An information processing method characterized in that the processing is executed by a computer.

[0162] (Appendix 9) Acquire a plurality of position data indicating the positions of each moving object in the target area at different times; generating first feature amount data representing spatial features of the area, the first feature amount data corresponding to a diagram representing timings at which one or more shapes of different types that can be formed by combinations of positions of moving objects on the area appear and disappear according to changes in resolution, based on the plurality of position data acquired by topological data analysis; learning a model that outputs a result of analyzing a risk due to a congestion situation in the area in accordance with input feature amount data that represents spatial characteristics of the area, based on the generated first feature amount data; An information processing device comprising a control unit. [Explanation of symbols]

[0163] 100 Information processing device 101 Location Data 110 Diagrams 111 First feature data 120 model 200 Information Processing Systems 201 Imaging device 202 Client device 210 Network 300,400 buses 301,401 CPU 302,402 memory 303,403 Network I / F 304,404 Recording media I / F 305,405 Recording media 306 Display 307 Input Device 406 Camera 500 storage section 501 Acquisition Department 502 Generation part 503 Learning Department 504 Analysis Department 505 Output section 600 point cloud data 610 2D pixel data 700 3D voxel data 800 point cloud 900, 1000, 1100 PD 1120,1210,1220,1230 PI 1300 point cloud dataset 1310,1320 feature vectors 1400 graphs 1500 total area 1501~1506 Individual area 1510 Output Screen

Claims

1. acquiring a plurality of position data indicating the positions of respective mobile objects in a region of interest at different times; generating first feature amount data representing spatial features of the area, the first feature amount data corresponding to a diagram representing timings at which one or more shapes of different types that can be formed by combinations of positions of moving objects on the area appear and disappear in accordance with a change in resolution, based on the plurality of position data acquired by topological data analysis; learning a model that outputs a result of analyzing a risk due to a congestion situation in the area in accordance with input feature amount data that represents spatial characteristics of the area, based on the generated first feature amount data; An information processing program that causes a computer to execute a process.

2. 2. The information processing program according to claim 1, wherein the moving object is a person.

3. 3. The information processing program according to claim 2, wherein the one or more shapes include at least one of a connecting shape, a ring shape, and a hollow shape that can be formed by a combination of positions of moving objects on the area.

4. The acquiring process includes: acquiring, for each of a plurality of time periods, a plurality of position data indicating the position of each of the moving objects in the area at different times within the time period; The generating process includes:

4. The information processing program according to claim 3, further comprising: generating, for each of the periods based on the acquired plurality of position data through topological data analysis, first feature data representing spatial characteristics of the region, corresponding to a diagram showing the timing at which each of the shapes appears and disappears depending on a change in resolution.

5. The learning process includes: The information processing program according to any one of claims 1 to 4, characterized in that the model is trained based on a combination of the generated first feature data and the presence or absence of risk due to congestion in the area during each of the periods.

6. generating second feature amount data representing spatial features on the area, the second feature amount data corresponding to a diagram representing timings at which the shapes appear and disappear according to changes in resolution, based on a plurality of position data representing positions of the respective moving objects on the area at different times within a predetermined period, by topological data analysis; outputting a result of analyzing the risk due to the congestion situation in the area during the predetermined period based on the generated second feature amount data using the trained model; 5. The information processing program according to claim 1, which causes the computer to execute processing.

7. acquiring a plurality of position data indicating the positions of respective mobile objects in a region of interest at different times; generating first feature amount data representing spatial features of the area, the first feature amount data corresponding to a diagram representing timings at which one or more shapes of different types that can be formed by combinations of positions of moving objects on the area appear and disappear in accordance with a change in resolution, based on the plurality of position data acquired by topological data analysis; learning a model that outputs a result of analyzing a risk due to a congestion situation in the area in accordance with input feature amount data that represents spatial characteristics of the area, based on the generated first feature amount data; An information processing method characterized in that the processing is executed by a computer.

8. acquiring a plurality of position data indicating the positions of respective mobile objects in a region of interest at different times; generating first feature amount data representing spatial features of the area, the first feature amount data corresponding to a diagram representing timings at which one or more shapes of different types that can be formed by combinations of positions of moving objects on the area appear and disappear in accordance with a change in resolution, based on the plurality of position data acquired by topological data analysis; learning a model that outputs a result of analyzing a risk due to a congestion situation in the area in accordance with input feature amount data that represents spatial characteristics of the area, based on the generated first feature amount data; An information processing device comprising a control unit.

Images