Information processing method, information processing apparatus, and information processing program
By generating and comparing heat maps to identify related regions and create predictive models, the method effectively predicts complex movements with reduced processing loads, facilitating real-time analysis and visualization.
Patent Information
- Application Number
- JP2023075774
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-05-01
- Publication Date
- 2024-11-14
AI Technical Summary
Existing methods struggle to predict complex movements such as traffic or human flow over a wide area with high accuracy while maintaining manageable processing loads, especially when real-time predictions are required and large-scale computers are not available.
The method involves creating multiple heat maps over time, comparing them to identify related regions, determining movement directions, and generating predictive models based on these regions to reduce processing loads.
This approach allows for accurate prediction of complex movements across wide areas with reduced processing demands, enabling real-time analysis and visualization of movement trends.
Smart Images

Figure 2024160598000001_ABST
Abstract
Description
[Technical field]
[0001] The present invention relates to an information processing method, an information processing device, and an information processing program. [Background technology]
[0002] To predict the flow of vehicles (traffic flow) in a small area, it is relatively easy to observe the number of vehicles at a branching point and calculate the results based on statistical methods or probability theory. For example, since vehicles basically travel on roads, possible events are limited to realistic patterns, making predictions relatively easy. Specifically, in the case of an intersection, excluding rare cases such as backing up and going off course, it is sufficient to predict three patterns: forward, right turn, and left turn.
[0003] On the other hand, if one were to try to perform predictions over a wide range with the same prediction accuracy, the number of branches that need to be estimated would be so large that the computational costs would be extremely high, and this is not something that can be easily achieved.
[0004] Patent Document 1 describes a method of calculating the flow of crowds in the imaging range of a fixed camera using a first surveillance image, calculating the distribution of object attributes in the imaging range of a mobile camera using a second surveillance image, and estimating the attribute distribution for a range not included in the imaging range of the mobile camera. Patent Document 1 also proposes a technology for estimating the flow of people by detecting the temporal movement of feature points of people, etc. in a camera image using optical flow, and for displaying the attribute distribution of people estimated from the camera image on a map in the form of a heat map. [Prior art documents] [Patent documents]
[0005] [Patent Document 1] Patent Publication No. 2022-89834 Summary of the Invention [Problem to be solved by the invention]
[0006] When predicting complex movements such as traffic flow and people flow over a wide area with the same prediction accuracy, it is possible to perform simulations over a long period of time, but in this case, it is difficult to apply the simulations to real-time predictions, etc. Also, it is possible to perform simulations using large-scale computers such as so-called supercomputers, but it is sometimes not possible to employ large-scale computers, and so there has been a demand for technology that reduces the processing load when predicting complex movements. The technology of Patent Document 1 also performs detailed analysis of images captured by a camera to make predictions, so there is a problem in that if predictions are to be made over a wide range with the same prediction accuracy, the amount of processing required becomes enormous.
[0007] The technology disclosed herein aims to provide a technology that reduces the processing load when estimating the movement of a moving object. [Means for solving the problem]
[0008] In order to solve the above problems, the information processing method of the present disclosure includes: creating a plurality of heat maps over time for each area on the map, the heat maps being assigned values according to at least one of the amount and attributes of the acquired moving objects; The heat maps created before and after the passage of time are compared, and regions that are related before and after the passage of time are extracted as related regions based on values of each region in the heat maps; Based on the position of the relevant region in the heat map before and after the passage of time, the movement direction and Extracting at least one of the movement information of the movement amount; predicting movement of the moving object based on the movement information; The control unit executes the above. Effect of the Invention
[0009] According to the technology of the present disclosure, it is possible to provide a technology that reduces the processing load when estimating the movement of a moving object. [Brief description of the drawings]
[0010] [Figure 1] FIG. 1 is a diagram showing the configuration of a movement prediction system according to this embodiment. [Diagram 2] FIG. 2 is a diagram illustrating a configuration of a movement prediction device. [Diagram 3] FIG. 3 is a diagram showing a map without heat map coloring. [Figure 4] FIG. 4 is a diagram illustrating an example of a heat map. [Diagram 5] FIG. 5 is a diagram showing heat maps before and after the passage of time. [Figure 6] FIG. 6 is a diagram showing the probability density distribution at points A to E. As shown in FIG. [Figure 7] FIG. 7 is a diagram showing an example of displaying the prediction result. [Figure 8] FIG. 8 is a diagram showing a process for generating a prediction model according to this embodiment. [Figure 9] FIG. 9 is a diagram showing a prediction process using the prediction model generated in the process of FIG. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0011] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. The configurations of the following embodiments are merely examples, and the present invention is not limited to the configurations of the embodiments.
[0012] First Embodiment FIG. 1 is a diagram showing a configuration of a movement prediction system 10 according to the present embodiment. The movement prediction system 10 includes a sensor 20 for detecting the amount and attributes of a moving body, and a movement prediction device 30 for predicting the movement of the moving body. The movement prediction device 30 is one form of an information processing device. In this embodiment, the moving body is, for example, a person, a vehicle, or an animal other than a human. The vehicle includes, for example, an automobile, a bicycle, a senior car, a wheelchair, and the like. Note that senior cars and wheelchairs are not included in the vehicles, and a person riding a senior car or a wheelchair may be considered as a moving body. In addition, the moving body may be a commodity moved by a person or a vehicle, or an object such as a drifting object in the sea, a lake, or a river. In the following embodiment, an example of predicting the movement of people (people flow) will be mainly described, but the same applies to the case of predicting the movement of vehicles (traffic flow) and the movement of goods (logistics). When predicting the movement of people in a wide area, the movement prediction system 10 creates a heat map based on the amount and attributes of people, compares the heat maps before and after the passage of time, extracts characteristic movements, models them, and uses them for prediction. In this way, the movement prediction system 10 of this embodiment grasps the flow of people by comparing heat maps and reduces the processing load. The technology of this embodiment can be applied to predicting movement in a wide area, and can predict a wide area such as, for example, the inside of a facility premises to an entire city, prefecture, or country without increasing the processing load too much. Hereinafter, the area that is the subject of this prediction is also referred to as the target area. This target area may not be a place where the movement direction and branching points are limited, such as a highway, but may be a site where mobile objects can move freely.
[0013] <Sensor> The sensor 20 detects the quantity and attributes of moving objects present in a target area. The sensor 20 may be, for example, a camera 21 that photographs the target area. The camera 21 transmits the photographed image (photographed image) to the movement prediction device 30 as moving object information. The movement prediction device 30 that receives the photographed image performs image analysis on the photographed image to determine the number (quantity) and attributes of people appearing in the photographed image, i.e., people present in the target area (photography location). The attributes that can be determined from the photographed image are The gender is information that can be determined from an image, such as walking, running, holding a cane, wearing glasses, wearing a hat, holding an umbrella, riding a bicycle, riding in a wheelchair, holding hands, or carrying a child. The camera 21 may be a fixed camera installed in a city or the like, a camera installed in a moving body such as a car, or a camera installed in a mobile terminal such as a smartphone or tablet terminal carried by a person. In particular, in order to detect information in the range to be predicted without omission, it is effective to install the camera 21 in a public transportation facility such as a route bus or a streetcar that runs a fixed route regularly. It is also effective to install a camera in a taxi that may run in various places in a city or the like. When the camera 21 moves with a car or a person, the camera 21 or a device connected to the camera 21 may transmit position information indicating the location where the image was taken together with the captured image to the movement prediction device 30 as moving body information.
[0014] The sensor 20 may be a communication terminal 22 carried by a person, such as a smartphone, a smart watch, smart glasses (glasses), a tablet terminal, etc. Examples of attributes acquired from the communication terminal 22 include location information, movement history, home address, and home zip code.
[0015] Furthermore, the sensor 20 may be a reader 23 that reads an IC tag or a barcode carried by a person. Examples of the reader 23 include ticket gates for trains and buses, card readers for convenience stores and vending machines, and readers that read point cards for stores. Examples of attributes acquired by the reader 23 include the names of stations where a person gets on and off, stores used, and the person's ID.
[0016] <Movement prediction device> FIG. 2 is a diagram showing a configuration of the movement prediction device 30. The movement prediction device 30 is a computer having a control unit 332, a memory 333, an input / output IF (interface) 334, and a communication IF 335, which are connected to each other by a connection bus 331. The control unit 332 processes input information and outputs the processing result, thereby controlling the entire device. The control unit 332 is also called a CPU (Central Processing Unit) or an MPU (Micro-processing unit). The control unit 332 is not limited to a single processor, and may have a multiprocessor configuration. Also, the control unit 332 may have a multicore configuration having multiple cores in a single chip connected by a single socket.
[0017] The memory 333 includes a main storage device and an auxiliary storage device. The main storage device is used, for example, as a working area for the control unit 332, a storage area for temporarily storing information processed by the control unit 332, and a buffer area for communication data. The main storage device is a storage medium for the control unit 332 to cache programs and data and to expand the working area. The main storage device includes, for example, a random access memory (RAM), a read only memory (ROM), and a flash memory. The auxiliary storage device is a storage medium for storing programs executed by the control unit 332, data used in information processing, setting information for operations, and the like. The auxiliary storage device is, for example, a hard-disk drive (HDD), a solid state drive (SSD), an erasable programmable ROM (EPROM), a flash memory, a USB memory, a memory card, and the like. .
[0018] The input / output IF 334 is an interface that inputs and outputs data between the movement prediction device 30 and peripheral devices connected thereto. The input / output IF 334 inputs and outputs data between devices such as a disk drive that reads data from a storage medium such as a CD or DVD, an operation unit, a display device, a microphone, a speaker, and a sensor. The operation unit is an input unit such as a mouse, keyboard, or touch panel that inputs information to the movement prediction device 30 by an operator's operation. The display device is an output unit that displays and outputs information such as processing results to the operator.
[0019] The communication IF 335 is an interface (communication module) that communicates with other devices via a communication line (network), and is also called a CCU (Communication Control Unit).
[0020] In the movement prediction device 30, the control unit 332 executes an application program, whereby the control unit 332 functions as each processing unit, such as the moving object information acquisition unit 301, the heat map generation unit 302, the area extraction unit 303, the feature estimation unit 304, the prediction model generation unit 305, the movement prediction unit 306, and the output control unit 307. That is, the control unit 332 can be used as each processing unit depending on the software being executed. However, it should be noted that some or all of the above processing units may be formed of hardware, such as a dedicated LSI (large scale integration) such as a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), or an FPGA (Field-Programmable Gate Array), a logic circuit, or other digital circuit. In addition, at least a part of each of the processing units may include an analog circuit. The control unit 332 may be configured such that one processor functions as multiple processing units, or may be configured to include multiple processors that function as one processing unit.
[0021] The mobile object information acquisition unit 301 acquires mobile object information indicating the quantity and attributes of the mobile object from the sensor 20. The mobile object information acquisition unit 301 acquires a captured image by connecting to, for example, the camera 21 via the communication IF 335. The mobile object information acquisition unit 301 also acquires mobile object information by connecting to, for example, the communication terminal 22 or the reader 23 via the communication IF 335. The mobile object information acquisition unit 301 may also acquire conditions such as season and weather from an information providing server together with the mobile object information. When acquiring mobile object information from the camera 21 or the communication terminal 22 that moves together with the mobile object, position information indicating the position of the camera 21 or the communication terminal 22 is acquired from the camera 21 or the communication terminal 22, so that it is possible to identify which area the mobile object information is in.
[0022] The heat map generating unit 302 creates a heat map in which a value is assigned according to the amount and attributes of the acquired mobile objects for each area on the map corresponding to the target area based on the mobile object information acquired by the mobile object information acquiring unit 301. FIG. 3 is a diagram showing a map 51 in a state where the heat map is not colored. The map 51 in FIG. 3 has coordinates corresponding to the target area, and for example, the longitude and latitude of the target area correspond to the X and Y coordinates of the map 51. The map 51 is also divided into a plurality of areas 52. In other words, the map 51 is composed of a plurality of areas (unit areas) 52, with one area 52 being one unit (one section).
[0023] Here, the heat map is, for example, a visualized graph in which the individual values of data (two-dimensional matrix data) in each two-dimensional region are expressed as colors or shades. In a general heat map, the values of each region are expressed as colors or shades so that they are easy for people to understand. In the present disclosure, the heat map is also used for computer judgment, so the heat map in this case includes the individual values (data values such as numerical values indicating colors or shades) before being expressed as colors or shades that are held in two dimensions (retained in each two-dimensional region). The values of the heat map described above are not limited to the values of raw data indicating information on quantity or attributes, but may be values converted to convert the values of the raw data into gradation data indicated by a predetermined number of stages (number of gradations). In this case, for example, gradation data is created by gradating the raw data so that values of quantities or attributes close to each other become one gradation value. In this way, it is easier to judge the relevance based on the data after it has been converted into gradation data than to judge the relevance from heat maps that are before and after in time based on raw data indicating information on quantity or attributes, and the processing load can be reduced. In particular, by converting multiple data such as quantities and multiple attributes into one gradation data (data with the same index / scale), judgments (for example, judging whether each value is close or far when judging grouping or correlation) become easier. Note that heat maps are not limited to two dimensions, but can be displayed in N dimensions (three or more dimensions). For example, the present invention may be applied to predict the movement of a moving object in a facility having multiple floors. In some cases, the heat map may be configured with three-dimensional data. Also, a plurality of two-dimensional heat maps may be prepared for each hierarchical level to predict movement between hierarchical levels.
[0024] The heat map generating unit 302 generates a heat map by assigning values to each area 52 on the map 51 according to the amount (number of people) of moving bodies and attributes. FIG. 4 is a diagram showing an example of a heat map 53. The heat map 53 shows the difference in values in each area 52 by different colors. Note that in FIG. 4, for convenience, the difference in color is shown by different hatching. For example, the shade of color is displayed according to the number of people or the number of vehicles. Also, the color may be displayed according to the value of the attribute. For example, local residents are displayed in red, city residents in green, and non-city residents in blue. Note that when multiple people with different attributes exist in one area 52, for example, when there are local residents and non-city residents, red and blue may be mixed to display as purple.
[0025] In addition, when displaying colors using RGB codes, attribute values (e.g., local residents, city residents, out-of-city residents) may be assigned to R, G, and B, respectively, and colors may be displayed using RGB codes with the number of people with these attributes as the gradation value. That is, if the number of local residents is large, the color is dark red, and if the number of local residents is small, the color is light red. In addition, instead of assigning three attributes to RGB, if each RGB has an 8-bit value, six attributes may be assigned by dividing it into 4-bit intervals. In addition, attribute and quantity values may be assigned to the shading value, and both the attribute and the quantity may be represented by the shading alone. In addition, the value of each area of the heat map is not limited to one that indicates both the quantity and attribute of the moving object, but may be a value that indicates at least one of the quantity and attribute. In addition, the value of each area of the heat map may be a value that indicates at least one of the color and shading of each area.
[0026] A heat map 53 in which each unit area 52 is colored in this way can be considered as an image in which each unit area 52 is a pixel. For example, a heat map 53 of a 12.5 km square area divided into 125 m square areas 52 can be considered as an image of 100 x 100 pixels.
[0027] The heat map generating unit 302 generates a heat map 53 every time a predetermined time (unit time) elapses. That is, the heat map generating unit 302 generates a plurality of heat maps 53 as time elapses.
[0028] The region extraction unit 303 compares the heat maps 53 before and after the passage of time among the multiple heat maps 53, and extracts the regions 52 that are related before and after the passage of time as related regions based on the values of each region 52 in the heat map 53. Fig. 5 is a diagram showing the heat maps 53 before and after the passage of time. In Fig. 5, reference numeral 531 indicates the heat map before the passage of time, and reference numeral 532 indicates the heat map after the passage of time. Note that the number of heat maps 53 to be compared is not limited to two, and may be three or more.
[0029] The region extraction unit 303 extracts, as related regions, regions 52 in which the color or shading information of the regions 52 (i.e., the amount or attribute of moving bodies) is similar and which are close to each other in the heat map 53 before and after the passage of time. Before associating these regions, grouping may be performed. In this grouping, when the values of adjacent regions 52 in one heat map 53 are similar, these regions are grouped, that is, a plurality of regions are consolidated into one large region. When extracting related regions in the heat map 53 before and after the passage of time, related regions may be extracted based on the size after grouping (the number of regions) and the shape of the region after grouping in addition to the value (color or shading) of each region. Here, the case where the values of the regions 52 are similar may include the case where the values of the regions 52 are completely the same, as well as the case where the number of people with the same attribute is close (for example, the difference in the number of people is within a predetermined value). For example, In the previous heat map 531, reference numerals 533 and 534 indicate groups of regions 52.
[0030] In addition, when the heat map 53 represents the quantity and attributes of moving bodies by color and shading, the region extraction unit 303 may group not only regions that are similar in both color and shading, but also regions that are similar in either color or shading. For example, when attributes are represented by color and quantities by shading, if the colors (attributes) are similar, even regions with distant shading (quantities) may be grouped. In addition, the quantity may be grouped only when it is equal to or greater than a threshold. For example, among regions where the quantity of moving bodies is equal to or greater than a threshold, regions where the quantity or attributes of moving bodies are similar may be grouped. In addition, the grouping conditions may be set so that, if the quantity of moving bodies is equal to or greater than a threshold, the region is grouped even if the quantity value is not close to that of an adjacent region.
[0031] For example, a group 533 in a heat map 531 before the passage of time and a group 535 in a heat map 532 after the passage of time have similar sizes after grouping and values of each region, and the region extraction unit 303 extracts these groups 533 and 535 as related regions. In addition, the region extraction unit 303 extracts, for example, a region 536 in the heat map 531 before the passage of time and a region 537 in the heat map 532 after the passage of time as related regions. Note that even if the values of the regions 52 are similar, if there is no change in position between the heat maps 531 and 532 before and after the passage of time, or if the number of moving bodies (e.g., the number of people) is less than a predetermined number, these regions 52 do not need to be extracted as related regions.
[0032] Feature estimation unit 304 estimates at least one of the movement direction and the movement amount as a feature amount based on the position of the related region in heat map 531 before the time has elapsed and the position of the related region in heat map 532 after the time has elapsed. In the example of Fig. 5, a movement vector indicated by arrow 541 is estimated from the position of related region (group) 533 and the position of related region (group) 535.
[0033] The prediction model generating unit 305 generates a prediction model indicating the probability of movement between the related regions 533 and 535 based on at least one of the extracted movement direction and movement amount (movement distance). For example, when the position of the related region for which the feature amount estimating unit 304 estimated the movement vector is from point A to point E, the prediction model generating unit 305 generates a prediction model by regarding which point the moving body moves to per unit time as a random variable. FIG. 6 is a diagram showing the probability density distribution at points A to E. The estimation of the probability density distribution in this prediction model can be realized by various methods such as Bayesian estimation, a mixed distribution model, and a K-NN method. Note that when creating a model, a plurality of data regarding the movement direction and movement amount of the same place (the same region of the map) may be collected, and the model may be created based on the plurality of data (data at a plurality of time points). For example, the movement direction and movement amount from the same region to multiple regions (destination regions) may be obtained for each destination region, and the model may be created based on the information for each of these multiple destination regions. When generating a prediction model, a model may be created for each condition, such as day of the week, time of day, season, weather, etc.
[0034] The movement prediction unit 306 predicts the movement of the moving objects based on a prediction model. For example, the current amount and position of the moving objects are input into the random variables of the prediction model, and the amount and position of the moving objects after a predetermined time are predicted. For example, the movement prediction unit 306 selects a prediction model suitable for prediction from a group of prediction models created in the past based on conditions such as the place to be predicted, the day of the week, the time period, the season, and the weather, and performs prediction based on the current amount and position of the moving objects using the selected prediction model.
[0035] In addition, the movement prediction unit 306 may predict the position after a predetermined time from the past movement direction and movement amount without using a prediction model. For example, when predicting the position after a predetermined time in real time, if there is a related area (group) that has moved in the same direction and movement amount at consecutive timings in the past, the related area (group) will continue to move in the same direction and movement amount even after the predetermined time. This method of prediction is particularly effective when predicting the movement of a moving object within a facility where the movement path (road, etc.) is not clearly defined (movement where the movement direction is not influenced by the shape of the road).
[0036] The output control unit 307 outputs the prediction result predicted by the movement prediction unit 306 to a display device or the like. Fig. 7 is a diagram showing a display example of the prediction result. In the example of Fig. 7, an arrow 541 indicating a movement vector and a message 56 based on the prediction result are displayed on a map 62 corresponding to the target area of the prediction. Note that the message 56 is a message corresponding to the prediction result that is registered in advance in a database, and is read out from the database and output according to the prediction result of the movement prediction unit 306.
[0037] <Movement prediction method> Next, a movement prediction method (information processing method) executed by the control unit 332 of the movement prediction device 30 will be described. Fig. 8 is a diagram showing a process for generating a prediction model according to this embodiment. When the movement prediction device 30 is powered on or when a start operation is performed by the user, the control unit 332 executes the process of Fig. 8 according to an application program. Note that the control unit 332 repeatedly executes the process of Fig. 8 until the movement prediction device 30 is powered off or until a termination operation is performed by the user.
[0038] In step S10, the control unit 332 acquires moving object information indicating the amount and attributes of moving objects from the sensor 20. For example, the control unit 332 acquires the moving object information from the camera 21, the communication terminal 22, or the reader 23.
[0039] In step S20, the control unit 332 generates a heat map for each area on the map corresponding to the target region, with values according to the amount and attributes of the moving objects acquired in step S10, and stores the heat map in the memory 333. The control unit 332 repeats the process of Fig. 8 for each unit time, thereby generating multiple heat maps over time.
[0040] In step S30, the control unit 332 compares the heat maps before and after the passage of time among the multiple heat maps generated in step S20, and extracts areas that are related before and after the passage of time as related areas based on the values of each unit area in the heat map 53.
[0041] In step S40, the control unit 332 estimates at least one of the movement direction and the movement amount as a feature amount based on the position of the related region in the heat map before the time has elapsed and the position of the related region in the heat map after the time has elapsed.
[0042] In step S50, the control unit 332 generates a prediction model indicating the probability of movement between related regions based on at least one of the estimated movement direction and the movement amount (movement distance).
[0043] Fig. 9 is a diagram showing a prediction process using the prediction model generated in the process of Fig. 8. In step S100, the control unit 332 acquires current moving object information from the sensor 20. Furthermore, when conditions are used when making the prediction, the control unit 332 acquires conditions such as the current season and weather.
[0044] In step S110, the control unit 332 inputs the moving object information acquired in step S100 into the prediction model to obtain a prediction result. If conditions are used for the prediction, the conditions acquired in step S100 are input.
[0045] In step S120, the control unit 332 outputs the prediction result obtained in step S110.
[0046] The result predicted by the movement prediction device 30 in this way is displayed as an arrow 541 indicating a movement vector and a message 56 based on the prediction result on a map 62 corresponding to the target area of the prediction, as shown in Fig. 7. This allows the user of the movement prediction device 30 to understand what kind of movement is predicted in the target area.
[0047] What is displayed on the display device may be not only the prediction result as shown in FIG. 7, but also the heat maps in the middle of the prediction as shown in FIG. 4 or FIG. 5. Also, as shown in FIG. 5, heat maps for each predetermined time may be displayed side by side, or the heat maps for each predetermined time may be switched and displayed according to the passage of time or a user operation. In this case, the grouped area may be highlighted. Note that the highlighting may be, for example, by changing the brightness or color tone of the grouped area, blinking it, drawing a frame of a specific color around the outer edge of the group, or the like.
[0048] The display device may be a display installed in a facility such as a town or a station and viewable by an unspecified number of people. The display device may also be installed in a target area of prediction and display prediction results to people or vehicle drivers moving in the target area.
[0049] The device for predicting movement (movement prediction device 30) and the device for displaying and operating the prediction result (management device) may be configured separately, and the movement prediction device 30 and the management device may be configured to be connected via a communication line. In this case, the movement prediction device 30 may output the prediction result to the management device, and the management device may display the prediction result on a display device provided in the management device, or the prediction result may be displayed on a display device installed in the city. In addition, the management device may present the received prediction result to an operator or AI, and execute measures (e.g., measures to alleviate congestion) according to the prediction result by the operator or AI. For example, when movement from a source area to a destination area is predicted and congestion occurs in the destination area, the occurrence of congestion in the destination area (prediction result) may be displayed on a display device provided in the source area to encourage the moving object to avoid the congestion.
[0050] Furthermore, the movement prediction device 30 is not limited to predicting movement on a plane, and may predict the movement of a moving object in a facility consisting of multiple floors. In this case, the movement prediction device 30 may, for example, construct a heat map using three-dimensional data, or have multiple heat maps for each floor, and extract areas that are related in the vertical direction (between floors) when comparing heat maps that are previous and next in time as related areas. Then, the movement prediction device 30 may extract movement information of at least one of the movement direction and the movement amount based on the position of this related area, and predict the movement in the vertical direction based on this movement information.
[0051] <Effects of the embodiment> The movement prediction device 30 of this embodiment creates multiple heat maps over time, compares the created heat maps before and after the time has elapsed, and extracts areas that are related before and after the time has elapsed as related areas based on the values of each area in the heat maps.The movement prediction device 30 then extracts movement information of at least one of the movement direction and movement amount based on the positions of the related areas in the heat maps before and after the time has elapsed, and predicts the movement of the moving object based on the movement information.
[0052] In this way, the movement prediction device 30 of this embodiment compares heat maps and predicts the movement direction and movement amount according to the positions of related areas before and after the passage of time, so there is no need to perform a process such as simulating all possible directions of movement in each area, and the processing load can be reduced. Therefore, it is possible to predict complex movements that do not depend on a specific route. In addition, the movement prediction device 30 of this embodiment predicts at least one of the movement direction and movement amount based on the positions of related areas in the heat maps before and after the passage of time, making it possible to perform conductor analysis of the moving body.
[0053] In addition, by appropriately adjusting the size (resolution) of the area that makes up the heat map, it is possible to set a balance between the prediction range, real-time performance, and accuracy that is appropriate for the purpose.
[0054] When extracting related regions, the movement prediction device 30 groups adjacent regions on the same heat map and extracts related regions based on the size after grouping in addition to the value of each region. This makes it possible to predict the tendency of movement across multiple regions.
[0055] The movement prediction device 30 generates a prediction model indicating the probability of movement between regions based on at least one of the extracted movement direction and movement amount. This reduces the processing load when predicting movement, making it easier to perform real-time prediction.
[0056] The movement prediction device 30 displays the prediction result on a display device. This makes it possible to present the prediction result in an easy-to-understand manner. For example, by displaying the prediction result superimposed on a map, the movement of people, cars, and objects over a wide area can be visualized in an easy-to-understand manner.
[0057] The movement prediction device 30 creates a heat map for each area on the map corresponding to a site where the moving object can freely move, with a value assigned to the moving object present in the site. The movement prediction device 30 also creates a heat map for each area on the map corresponding to a wide area such as an entire city, prefecture, or country, with a value assigned to the moving object present in the area. Even when predicting the movement of a complex moving object that does not depend on a specific route in this way, the movement prediction device of this embodiment can easily perform the prediction by making a prediction based on a comparison of heat maps according to the amount and attributes of the moving object.
[0058] <Other> The above-described embodiment is merely an example, and addition, omission, substitution, combination, and other modifications of the configuration are possible as appropriate without departing from the spirit of the present disclosure.
[0059] The technology of the present disclosure may also be an information processing program for causing a control unit to execute the above-mentioned movement prediction method (information processing method). Furthermore, the technology of the present disclosure may also be a non-temporary recording of the information processing program on a computer-readable recording medium. The computer can provide the function by reading and executing the program on the recording medium.
[0060] Here, a computer-readable recording medium refers to a recording medium that stores information such as data and programs electrically, magnetically, optically, mechanically, or chemically and can be read by a computer. Among such recording media, those that can be removed from a computer include, for example, flexible disks, magneto-optical disks, CDs (Compact Discs), CD-R / Ws, DVDs (Digital Versatile Disks), Blu-ray Discs (Blu-ray (registered trademark) Discs), DAT, 8 mm tapes, memory cards such as flash memory, etc. Furthermore, examples of recording media that are fixed to a computer include hard disks and ROMs (read-only memories). [Explanation of symbols]
[0061] 10: Movement prediction system 20: Sensor 21: Camera 22: Communication terminal 23: Leader 30: Movement prediction device 51: Map 52: Area (unit area) 53: Heat Map 56: Message 62: Map 301: Mobile information acquisition unit 302: Heat map generation unit 303: Area extraction part 304: Feature estimation unit 305: Prediction model generation unit 306: Movement prediction unit 307: Output control section 331: Connection bus 332: Control unit 333: Memory 531,532: Heatmap 533~535: Group
Claims
1. creating a plurality of heat maps over time for each area on the map, the heat maps being assigned values according to at least one of the amount and attributes of the acquired moving objects; The heat maps created before and after the passage of time are compared, and regions that are related before and after the passage of time are extracted as related regions based on values of each region in the heat maps; Extracting movement information of at least one of a movement direction and a movement amount based on the position of the relevant region in the heat map before and after the passage of time; and predicting the movement of the moving object based on the movement information. The control unit executes the information processing method.
2. The value for each area on the map in the heat map is a value that represents at least one of the color and the shade of each area. The information processing method according to claim 1 .
3. The control unit: When extracting the related regions, adjacent regions on the same heat map are grouped, and related regions are extracted based on the size after grouping in addition to the value of each region.
3. The information processing method according to claim 1 or 2.
4. The control unit: A prediction model indicating a probability of movement between the regions is generated based on the extracted movement information.
3. The information processing method according to claim 1 or 2.
5. The control unit: Predicting the movement of the moving object based on the prediction model The information processing method according to claim 4.
6. The control unit: The result of the prediction is displayed on a display device. The information processing method according to claim 5.
7. The control unit: A heat map is created for each area on the map corresponding to a site where the mobile object can freely move, with a value assigned to the mobile object present in the site.
3. The information processing method according to claim 1 or 2.
8. A control unit extracts movement information used to predict the movement of a moving object, The control unit is creating a plurality of heat maps over time for each area on the map, the heat maps being assigned values according to at least one of the amount and attributes of the acquired moving objects; The heat maps created before and after the passage of time are compared, and regions that are related before and after the passage of time are extracted as related regions based on values of each region in the heat maps; extracting the movement information, which is at least one of a movement direction and a movement amount, based on the positions of the relevant regions in the heat map before and after the passage of time; Information processing device.
9. Information processing to be executed by a computer to extract movement information to be used for predicting the movement of a moving object A program, creating a plurality of heat maps over time for each area on the map, the heat maps being assigned values according to at least one of the amount and attributes of the acquired moving objects; The heat maps created before and after the passage of time are compared, and regions that are related before and after the passage of time are extracted as related regions based on values of each region in the heat maps; extracting the movement information, which is at least one of a movement direction and a movement amount, based on the positions of the relevant regions in the heat map before and after the passage of time; Information processing program.
Citation Information
Patent Citations
Video monitoring system, video monitoring method, and program
JP2022089834A
Cited By
Multi-radio communication system and communication control device
WO2026191109A1
