JP2019012494A - Human flow estimating device, human flow estimating system and human flow estimating method - Google Patents

Abstract

To provide a human flow estimating device that makes it possible to set even an open space as a target area for human flow estimation, and performs highly accurate human flow estimation based on measurement data from a plurality of human flow measurement devices having different measurement characteristics.SOLUTION: In the human flow estimating system, the human flow estimating device 10 includes: an analysis unit 107 for acquiring a number of passes and traveling directions of pedestrians from two or more kinds of human flow measurement devices having different measurement characteristics, generating, as hypothesis OD data, plural pieces of OD data, which is a number of pedestrians heading from an actual entrance to an exit or a number of pedestrians heading from a virtual entrance to a virtual exit, and performing human flow analysis; a data assimilation OD estimation unit 108 for integrating the number of passes and the traveling directions of the pedestrians acquired from the human flow measurement devices with the data obtained from the analysis unit based on an allowable measurement error determined on the basis of the human flow measurement devices to generate estimated OD data; and a display control unit 110 for displaying the estimated OD data obtained from the data assimilation OD estimation unit 108 on a display unit 13 as a human flow estimation result.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a human flow estimation device, a human flow estimation system, and a human flow estimation determination method for estimating a human flow (human flow) in a target area.

  In recent years, attention has been focused on reducing urban congestion and ensuring the safety of pedestrians. Specifically, there is a concern that the concentration of tourists in urban areas may lead to dangerous accidents such as delays in arrival and departure of trains at stations and, in some cases, falling from homes. In addition, when an event is held, there may be a problem that a pedestrian cannot pass safely due to irregular congestion around the venue. Therefore, in order to ensure the safety and comfort of pedestrians, it is necessary to grasp the current state of congestion and predict the near future, control the flow of people by guidance and guidance, and reduce congestion. .

  In order to solve the above problems, for example, a technique described in Patent Document 1 has been proposed. Patent Document 1 discloses a congestion degree estimation device for estimating the degree of congestion such as whether the congestion is normal or abnormal in consideration of current and past train operation statuses and change patterns of human movements for station premises. ing. In a boarding / exiting facility such as a station premises, it is possible to estimate a person's movement status using a predetermined flow calculation method based on the combined number of entrance and exit (hereinafter referred to as OD: Origin Destination) data obtained by a ticket gate. It becomes. In a target area where an OD data acquisition unit such as a ticket gate is installed, it is possible to accurately estimate a person's movement status.

  A technique described in Patent Document 2 is also known. In Patent Document 2, in order to perform a large-scale human flow distribution estimation process more efficiently, a change in human flow is expressed using a distributed platform. This distributed platform generates a large number of hypotheses based on the human flow model obtained from the 100% survey results, and can evaluate the compatibility with the measurement results from the human flow measurement device to calculate the human flow distribution consistent with the measurement results. is there.

JP 2012-118790 A JP 2014-112292 A

However, in the technique disclosed in Patent Document 1, when a space (hereinafter referred to as an open space) where there is no entrance gate such as in a city or a commercial facility is targeted, it is difficult to obtain OD data, and the movement situation of people is difficult. Cannot be estimated. Even if an image acquisition device such as a camera is used, information on the number of people and the direction of travel can be obtained. However, obtaining OD data that associates the target person from the entrance to the exit, such as where to go, This is difficult unless pedestrians are identified and tracked.
In addition, in the technique described in Patent Document 2, in order to obtain a total survey result, the configuration is obtained by combining a preliminary survey based on a questionnaire and GPS information of a mobile phone, but in the first place a total survey result is obtained. Is difficult.

  Therefore, the present invention can be a target area for human flow estimation even in an open space, and can enable high-precision human flow estimation based on measurement data from a plurality of human flow measurement devices having different measurement characteristics. A human flow estimation device, a human flow estimation method, and a human flow estimation system are provided.

In order to solve the above problems, the human flow estimation device according to the present invention acquires the number of passing people and the direction of travel of pedestrians acquired from two or more types of human flow measurement devices having different measurement characteristics, and from the actual entrance to the exit. Based on the human flow measurement device, an analysis unit for generating a plurality of OD data as hypothetical OD data, which is the number of pedestrians heading to or the number of pedestrians heading from a set virtual entrance to a virtual exit, A data assimilation OD estimation unit that generates estimated OD data by integrating the number of passing people and the pedestrian travel direction acquired from the human flow measurement device and the data obtained from the analysis unit based on the allowable measurement error determined; And a display control unit that displays the estimated OD data obtained from the data assimilation OD estimation unit on a display unit as a human flow estimation result.
In addition, the human flow estimation system according to the present invention includes two or more types of human flow measurement devices having different measurement characteristics, and human flow estimation that performs human flow estimation based on at least the number of passing people and the direction of travel of the pedestrian measured by the human flow measurement device. The human flow estimation device obtains the number of passing people and the traveling direction of the pedestrians measured by the human flow measuring device, and the number of pedestrians moving from the actual entrance to the exit or a set virtual number Based on the permissible measurement error determined based on the human flow measurement device, an analysis unit that generates a plurality of OD data that is the number of pedestrians going from the entrance to the virtual exit as hypothetical OD data, and the human flow measurement A data assimilation OD estimation unit that generates estimated OD data by integrating the number of passing people and the direction of travel of pedestrians acquired from the device and the data obtained from the analysis unit; And having a display control unit for displaying on the display unit the estimated OD data obtained from the data assimilation OD estimating section as pedestrian flow estimation result.
Further, the human flow estimation method according to the present invention has at least two types of human flow measurement devices having different measurement characteristics, and estimates the human flow based on the number of passing people and the direction of travel of pedestrians measured by the human flow measurement device. A human flow estimation method to be performed, wherein the number of passing people and the pedestrian traveling direction measured by the human flow measuring device are acquired, and the number of pedestrians heading from an actual entrance to an exit or a virtual entrance to be set is virtual. A plurality of OD data, which is the number of pedestrians heading to the exit, is generated as hypothetical OD data, and a human flow analysis is performed. Based on an allowable measurement error determined based on the human flow measurement device, the number of passing people acquired from the human flow measurement device Also, the estimated OD data is generated by integrating the pedestrian direction and the data obtained from the result of the human flow analysis, and the generated estimated OD data is used as the human flow estimation result. And displaying on the display unit.

According to the present invention, even if it is an open space, it can be set as a target area for human flow estimation, and high-precision human flow estimation can be performed based on measurement data from a plurality of human flow measurement devices having different measurement characteristics. It is possible to provide a human flow estimation device, a human flow estimation method, and a human flow estimation system.
Problems, configurations, and effects other than those described above will be clarified by the following description of embodiments.

BRIEF DESCRIPTION OF THE DRAWINGS It is a whole schematic block diagram of the human flow estimation system of Example 1 which concerns on one Example of this invention, Comprising: It is a functional block diagram of a human flow estimation apparatus. It is a flowchart which shows the process of the human flow estimation apparatus shown in FIG. It is a figure which shows the measurement position and measurement range in an example of an analysis model. It is a figure which shows OD data table. It is a schematic diagram showing the example of a data processing regarding the advancing direction of a pedestrian. It is a figure which shows the data structure of the data acquired from the measurement part stored in the memory | storage part which comprises the human flow estimation apparatus shown in FIG. It is a conceptual diagram of the human flow analysis by the analysis part which comprises the human flow estimation apparatus shown in FIG. It is a conceptual diagram of the data assimilation performed in the data assimilation OD estimation part which comprises the human flow estimation apparatus shown in FIG. It is an arrangement schematic diagram of a human current measuring device in a facility. It is a screen display example of the human flow estimation result displayed on the display part which comprises the human flow estimation apparatus shown in FIG. It is a whole schematic block diagram of the human flow estimation system of Example 2 which concerns on the other Example of this invention. It is a functional block diagram of the human flow estimation apparatus which comprises the human flow estimation system shown in FIG. It is a figure which shows the analysis model at the time of making the city which is a completely open space into the object area | region of human flow estimation. 12 is a screen display example of a human flow estimation result displayed on a display unit and a portable terminal that constitute the human flow estimation device shown in FIG. 11.

In this specification, “OD data” refers to the number of pedestrians heading from the actual entrance to the exit, or from a virtual entrance set to a virtual exit when a completely open space is the subject of human flow estimation. The number of pedestrians heading. The actual entrance and the set virtual entrance may be referred to as a departure place, and the actual exit and the set virtual exit may be referred to as an arrival place.
Embodiments of the present invention will be described below with reference to the drawings.

<Overall configuration of human flow estimation system 1>
FIG. 1 is an overall schematic configuration diagram of a human flow estimation system according to a first embodiment of the present invention, and is a functional block diagram of a human flow estimation device. As shown in FIG. 1, the human flow estimation system 1 includes two or more types of first human flow measurement devices 101a, second human flow measurement devices 101b, Nth human flow measurement devices 101n, and human flow estimation devices 10 having different measurement characteristics. Is done. Here, the first human flow measurement device 101a, the second human flow measurement device 101b, and the Nth human flow measurement device 101n constitute the measurement unit 102, and N ≧ 2. The number of human flow measurement devices is not limited to three, ie, the first human flow measurement device 101a, the second human flow measurement device 101b, and the Nth human flow measurement device 101n, and is appropriately set according to the target area of the human flow measurement. What is necessary is just to determine the number of person flow measuring devices. Moreover, as long as it has two or more types of human flow measurement devices having at least different measurement characteristics, it is not necessary that all the human flow measurement devices have different measurement characteristics. The human flow measurement device has a function of measuring the number of people in the measurement area and the traveling direction data of the pedestrian. In addition, two or more types of human flow measurement devices having different measurement characteristics are, for example, devices that have a wide measurement range but are difficult to install in an environment with many shields. Although the measurement range is narrow, it can be installed even in an environment with high accuracy and many shields. Furthermore, the measurement range or the measurement sensitivity is different even if the measurement range is the same.

<Configuration of the human flow estimation device 10>
As shown in FIG. 1, the human flow estimation device 10 has an input unit that allows a user to input an allowable measurement error (details will be described later) based on specifications or an analysis model to be described in detail later for each of the human flow measurement devices 101a to 101n. 12, a processing unit 103, and a display unit 13. The processing unit 103 receives an allowable measurement error (details will be described later) input from the input unit 12 or an input I / F 104 that imports an analysis model to be described later in detail, a hypothesis OD generation unit 105, and an OD generated by the hypothesis OD generation unit 105. A human flow analysis unit 106 that performs human flow analysis based on data, a data assimilation OD estimation unit 108, a storage unit 109, a display control unit 110, and a communication I / F 111 are provided, which are connected to each other via an internal bus 112. The hypothesis OD generation unit 105 and the human flow analysis unit 106 constitute an analysis unit 107. The hypothesis OD generation unit 105, the human flow analysis unit 106, the data assimilation OD estimation unit 108, and the display control unit 110 are, for example, a processor such as a CPU, a ROM that stores a program, and a program that is read from the ROM. This is realized by a storage device such as a RAM that temporarily stores process data and the like.
The operation of each part of the human flow estimation device 10 will be described below.

<Operation of the human flow estimation device 10>
FIG. 2 is a flowchart showing processing steps of the human flow estimation apparatus 10 shown in FIG.

[Step S11: Initial setting]
As shown in FIG. 2, in step S11, as an initial setting, an allowable measurement error (details will be described later) based on the specifications is input via the input unit 12 for each of the analysis model and the human flow measurement devices 101a to 101n, and the input I / F 104 is input. Is taken into the processing unit 103. The taken analysis model and the permissible measurement error (details will be described later) based on the specifications of each of the human flow measurement devices 101 a to 101 n are stored in a predetermined storage area of the storage unit 109 via the internal bus 111 from the input I / F 104.

Here, the analysis model will be described. FIG. 3 is a diagram illustrating a measurement position and a measurement range in an example of the analysis model. In the following, an example of a method for modeling the congestion estimation target region 20 and a case where two types of human flow measurement devices 101a and 101b are used as the human flow measurement device constituting the measurement unit 102 will be described. The installation will be described as follows. In FIG. 3, there are three entrances indicated by entrance A201, entrance B202, and entrance C203, and pedestrians come and go in both directions at all entrances. The measurement area 204 of the first human flow measurement device represents an area that can be measured by the first human flow measurement device 101a. The measurement area 205 of the second human flow measurement device represents an area that can be measured by the second human flow measurement device 101b. The structure 206 represents an area through which a person cannot pass. In the measurement region 204 of the first human flow measurement device assumed in FIG. 3, for example, the first human flow measurement device 101a having a wide measurement range represented by a laser radar is applied in order to predict a rough human flow in the target region. . Laser radars are difficult to measure at locations with many shields or crowded people, so the measurement position is so limited that the area that can be installed is limited. On the other hand, the second person flow measurement apparatus 101b performs measurement using, for example, a camera. The measurement with a camera has the advantage that the measurement range per unit is narrow but the accuracy is high, and the measurement can be performed even in a place with many shields or in a congested environment. Therefore, it is preferable to apply to a complicated passage where a laser radar cannot be installed or an area with many people.
In the present embodiment, a case where a laser radar and a camera are used as the first human flow measurement device 101a and the second human flow measurement device 101b is shown, but the present invention is not necessarily limited thereto. For example, an infrared tag, GPS, or a human sensor may be used. The first human flow measurement device 101a, the second human flow measurement device 101b, and the Nth human flow measurement device 101n constituting the measurement unit 102 (in the example shown in FIG. 3, the first human flow measurement device 101a and the second human flow measurement device 101b). In addition, as long as at least two types of human flow measurement devices 101 to 103 are included, a large number of the same type of human flow measurement devices may be installed.
As described above, the analysis model of the congestion estimation target region 20 illustrated in FIG. 3 is input via the input unit 12 and is taken into the processing unit 103 by the input I / F 104.

[Step S12: Hypothesis OD generation]
Returning to FIG. 2, the hypothesis OD generation unit 105 in step S12 generates a hypothesis OD. Specifically, the hypothesis OD generation unit 105 accesses the storage unit 109 via the internal bus 112, reads the stored analysis model described above, and generates hypothesis OD data.
Here, the generated hypothesis OD data will be described. FIG. 4 shows an OD data table. FIG. 4 shows an example of OD data (hypothesis OD) estimated by the hypothesis OD generation unit 105 for the analysis model shown in FIG. As shown in FIG. 4, the generated OD data (hypothesis OD data) is 250 people from the entrance A (O_A) to the exit B (D_B), and from the entrance A (O_A) to the exit C (D_C). 150 people, 600 people going from entrance B (O_B) to exit A (D_A), 350 people going from entrance B (O_B) to exit C (D_C), exit A (from O_C) to exit A (O_C) Six hypothetical OD data are generated in which the number of people going to D_A) is 100 people and the number of people going from the entrance C (O_C) to the exit B (D_B) is 350 people. In other words, if the estimation accuracy is poor even in one of these six conditions (six hypothesis OD data), the overall human flow changes and the estimation accuracy deteriorates. Therefore, it is required to accurately estimate all of these OD data. To improve accuracy, the number of installed first person flow measurement devices 101a, second person flow measurement devices 101b, and Nth person flow measurement devices 101n is increased, and the measurement positions are determined. It can be improved by devising.

  Moreover, the schematic diagram showing the example of a data processing regarding the advancing direction of a pedestrian is shown in FIG. FIG. 5 shows an example in which data processing is performed for the traveling direction 40 of the pedestrian 401 from an image obtained by the second human flow measurement device 101b (camera) as the human flow measurement device constituting the measurement unit 102. Here, a case is shown in which an image processing algorithm that can detect the traveling directions 402 to 409 of the pedestrian 401 in eight directions is applied. For example, in a passage where the flow of pedestrians is limited to two directions, the upper left direction 402, the upper direction 403, the upper right direction 404, and the right direction 405 may be integrated and handled as the upper direction 403. Similarly, the lower right direction 406, the lower direction 407, the lower left direction 408, and the left direction 409 may be integrated and handled as the lower direction 407. Using such a method, the traveling direction is not limited to eight directions and two directions, and any direction or number can be applied. The number of directions handled by all measurement data does not need to be the same, and the directions may be selectively defined based on the flow of people in the target area.

  FIG. 6 shows a data structure of data acquired from the measurement unit 102 stored in the storage unit 109 constituting the human flow estimation device 10 shown in FIG. Here, in the 1st person flow measuring device 101a, the 2nd person flow measuring device 101b, and the Nth person flow measuring device 101n, it is obtained from the measured direction of the pedestrian and the number of passing people, the specification of the person flow measuring device or the preliminary examination, respectively. Are stored in association with each other. In the example shown in FIG. 6, in the congestion estimation target region 20 corresponding to the analysis model shown in FIG. 2 described above, data (traveling direction and number of people passing) measured between 8:30 and 9:30 are shown. .

The first person flow measurement apparatus 101a indicates that the number of passing people whose traveling direction is upward is 440 and the number of passing persons whose traveling direction is downward is 654. It shows that the allowable measurement error of 101a is 15%.
Regarding the second person flow measuring device 101b, the number of people passing in the upper left direction is 1, the number of people passing in the upward direction is 13, the number of people passing in the upper right direction is 25, and the direction of movement is right. The number of passing people is 4, the number of passing people is 10 in the lower right direction, the number of passing people is 32 in the moving direction, the number of passing people is 18 in the lower left direction, and the moving direction is left It is shown that the number of people passing through was 2 people, and that the allowable measurement error of the second person flow measurement device 101b was 5%.
As for the Nth person flow measuring device 101n, the number of people passing in the upward direction is 53, the number of people passing in the right direction is 75, the number of people passing in the downward direction is 113, and the direction of travel is left. This indicates that the number of people passing in the direction is 25, and that the allowable measurement error of the Nth human flow measuring device 101n is 10%.

  As the data structure of the data acquired from the measurement unit 102 stored in the storage unit 109, in addition to the measurement data of the direction of travel and the number of passing people, an allowable measurement error is stored in the storage unit 109 for each human flow measurement device. Thus, by simultaneously evaluating the allowable measurement error, it is possible to perform data integration in consideration of the difference in measurement characteristics among the first human flow measurement device 101a, the second human flow measurement device 101b, and the Nth human flow measurement device 101n. . In addition, as described above, it is not necessary to use the same number of handling directions in the first person flow measurement device 101a, the second person flow measurement device 101b, and the Nth person flow measurement device 101n, and each measurement is performed in consideration of the person flow. You may define by point. The handling of the allowable measurement error of the measurement data will be described later with the data assimilation method using FIG.

[Step S13: Data Acquisition from Human Flow Analysis Unit 106 and Measurement Unit 102]
Returning to FIG. 2, in step S <b> 13, the data assimilation OD estimation unit 108 acquires the number of persons and the traveling direction from the human flow analysis unit 106 and the measurement unit 102 via the internal bus 111. The measurement data of the number of persons and the traveling direction acquired from the measurement unit 102 is the data shown in FIG.
Here, the human flow analysis unit 106 constituting the analysis unit 107 will be described. FIG. 7 is a conceptual diagram of the human flow analysis by the analysis unit constituting the human flow estimation device shown in FIG. In the analysis model (congestion estimation target region 20) shown in FIG. 3 described above, the measurement region 204 of the first human flow measurement device and the measurement region 205 of the second human flow measurement device are limited regions in the analysis model. For this reason, it is important to supplement the human flow analysis based on the trace results in the human flow analysis unit 106 constituting the analysis unit 107. FIG. 7 shows an example of the movement process of the human flow analysis 60 performed by the human flow analysis unit 106, and the region of the human flow analysis 60 shown in FIG. 7 is the analysis model (congestion estimation target region) shown in FIG. 20) A part of the area is enlarged. The human flow analysis unit 106 is based on the hypothesis OD data generated by the hypothesis OD generation unit 105 constituting the analysis unit 107 described above, that is, based on the estimated OD data shown in FIG. Assuming that it exists, the movement of each pedestrian in the target area is derived by calculation.

  Here, as an example, a human flow calculation method for a target pedestrian 601 from the entrance 603 to the exit 604 will be described. In order to go from the entrance 603 to the exit 604, it is preferable to select the shortest route 606 from a plurality of route candidates 605. However, since the pedestrian 608 already exists on another shortest path 606 in the calculation grid 602 ahead of the shortest path 606, it is not possible to proceed. Therefore, a path 607 approaching the exit 604 can be selected from the adjacent calculation grids 602. By repeating the same operation, the human flow analysis unit 106 analytically estimates a human flow from the entrance 603 to the exit 604. Here, the analysis method is not limited to the above-described algorithm. For example, when a pedestrian 608 exists on the adjacent shortest path 606, the target pedestrian 601 moves until the pedestrian 608 on the shortest path 606 moves and the calculation grid 602 becomes free. You can also choose not to advance. Further, a method of estimating human flow by stochastically combining whether to advance or not progress may be used, and the calculation method is not limited.

[Step S14: Data Assimilation]
Returning to FIG. 2, in step S <b> 14, the data assimilation OD estimation unit 108 integrates data from the acquired results of the human flow analysis unit 106 constituting the measurement unit 102 and the analysis unit 107 in consideration of the number of people, the traveling direction, and the allowable error. OD data is estimated. A specific method of data assimilation will be described with reference to FIG. FIG. 8 is a conceptual diagram of data assimilation executed by the data assimilation OD estimation unit 108 constituting the human flow estimation device 10 shown in FIG. As shown in FIG. 8, data assimilation 70 is a technique for estimating an actual state with higher accuracy by combining analysis and measurement. Here, although the ensemble Kalman filter is applied as an example of the data assimilation 70, the present invention is not limited to this method, and a particle filter, a variational method, or the like may be applied. The ensemble Kalman filter is based on Bayesian estimation. If the predicted value by analysis is x and the measured value is y, the relationship of the following equation (1) is obtained by Bayes' theorem.

  Here, p (a) represents a probability density function of the state variable a. P (a | b) is a conditional probability density function of a when the value of b is fixed. P (x) on the right side is the probability density distribution 702 of the predicted human flow analysis value x, and p (y | x) is the probability density distribution 703 of the measured value. Here, the probability density distribution 703 of the measured values is assumed to be a normal distribution, the number of passing people is an average value, and the allowable measurement error is used to calculate variance. The larger the allowable measurement error, the larger the probability density distribution 703 is expressed. The P (x | y) on the left side represents the probability distribution of the analytical solution after data assimilation and is referred to as a posterior distribution 705. The posterior distribution 705 estimated by the data assimilation 70 is determined by the analysis predicted value 702 and the measured value 704, and can be converged to the true value 701 by repeating a series of calculations. By applying the numerical value of the OD data obtained by convergence to the processing of the human flow analysis unit 106 as definition OD data, it is possible to estimate the human flow.

FIG. 8 is a diagram for explaining assimilation processing of one OD data. The number of pedestrians from the OD data (from the entrance to the exit (six hypothesis OD data corresponding to the six conditions shown in FIG. 4)): 8 is generated for each hypothetical OD data) and processed until convergence to a true value 701 as shown in the lower diagram of FIG. 8 while avoiding mutual interference. In other words, for example, processing is performed until convergence is achieved while maintaining consistency among the six hypothesis OD data shown in FIG.
In the example shown in FIG. 8, a probability density distribution 703 of measured values when the second human current measuring device 101 b (camera) is used as the human current measuring device, and the one second human current measuring device 101 b ( This is a probability density distribution for image data captured by a camera. Therefore, by integrating the probability density distribution 702 of the predicted human flow analysis value shown in FIG. 8 and the measurement result by the second human flow measuring device 101b (camera), the analysis solution after data assimilation is shown as shown in the lower diagram of FIG. A posterior distribution 705 which is a probability distribution is obtained. Note that the data assimilation OD estimation unit 108 performs the same process on the measurement result obtained by the first human flow measurement apparatus 101a (laser radar).

[Step S15: Parameter Correction]
Returning to FIG. 2, in step S15, the data assimilation OD estimation unit 108 corrects the parameters. Specifically, in FIG. 8 described above, the “state variable” on the horizontal axis is the number of pedestrians. For example, the probability of the human flow analysis prediction value when the parameter that is the state variable is changed from 100 to 1000 people. A density distribution 702 and a probability density distribution 703 of measurement values are obtained.

[Step S16: Convergence Determination [DO]]
In step S16 shown in FIG. 2, the data assimilation OD estimation unit 108 obtains the estimated OD data (the posterior distribution 705 estimated by the data assimilation 70) obtained after the parameter correction as the state variable in the above step S15. It is determined whether or not it has converged to the true value 701. If it has not converged, the process returns to step S12, and the processes up to step S16 are repeated. On the other hand, when it determines with having converged in step S16, a process is complete | finished. In addition, the display on the display unit 13 of the human flow estimation result obtained after the process ends will be described later.

  In the present embodiment, the configuration in which the human flow estimation device 10 executes Steps S11 to S16 illustrated in FIG. 2 has been described. However, instead of this, the following processing may be executed. In other words, the OD data estimated by the data assimilation OD estimation unit 108 is compared with the OD data estimated in the previous step. Then, the hypothesis OD generation unit 105 generates OD data again. Here, the data assimilation result is reflected, and the range in which the hypothesis OD data is generated is gradually narrowed to the vicinity of the estimated value (the above-described true value 701). When the difference between the OD data estimated by the data assimilation OD estimation unit 108 and the OD data similarly estimated in the previous step becomes smaller than a threshold value, the estimated OD data is considered to have converged. The process ends.

<Display of human flow estimation results>
FIG. 9 is an arrangement schematic diagram of the human flow measurement device in the facility, and FIG. 10 is a screen display example of the human flow estimation result displayed on the display unit 13 constituting the human flow estimation device 10 shown in FIG.
FIG. 9 shows a schematic diagram in which two types of first person flow measurement devices 101a and second person flow measurement devices 101b having different measurement characteristics are installed in the facility. The first human flow measurement apparatus 101a is a human flow measurement apparatus with a wide viewing angle and a wide measurement range but low measurement accuracy. On the other hand, the second human flow measurement device 101b is a human flow measurement device that can measure with high accuracy although the viewing angle is narrow and the measurement range is narrow. Here, it is desirable that the number of installed first person flow measurement apparatuses 101a and second person flow measurement apparatuses 101b is at least larger than the estimated number of OD data.

The entrance / exit 801 to be studied in the analysis model shown in FIG. 9 is not limited to the open portion of the passage, but also includes a staircase (including an escalator) 802 or an elevator 803 where a pedestrian enters and exits. Based on the number of people information measured by the first person flow measuring device 101a and the second person flow measuring device 101b and the measurement data related to the direction of travel of the pedestrian, the person flow is estimated by the operation of the person flow estimating device 10 described above.
The human flow estimation result obtained by the data assimilation OD estimation unit 108 constituting the human flow estimation device 10 is displayed on the screen of the display unit 13 via the internal bus 111 and the display control unit 110. The screen display example shown in FIG. 10 shows the human flow estimation result displayed on the screen of the display unit 13.
Based on the result of the human flow analysis performed on the analysis model shown in FIG. 9, the congestion degree distribution 901 is expressed in shades, and the human flow estimation result is displayed on the human flow estimation result display screen 90. A black circle on the human flow estimation result display screen 90 represents the pedestrian 401. For example, in FIG. 10, it is estimated that congestion is occurring in the vicinity of the elevator 803, and the congested area 902 can be easily identified visually. From the result, a condition for reducing congestion by artificially changing the OD data is searched, and a crowd flow mitigation effect can be expected by displaying the human flow guidance display 903 based on the result. Note that the congestion mitigation method is not limited to the above method, and a method of analytically searching for conditions is also conceivable. In addition, if a display for guidance display is installed in the facility in advance, the flow guidance display 903 is a message or an arrow indicating the direction in which the guidance corresponding to the flow guidance display 903 is to be displayed on the display. An image should be displayed. Further, if a guide light is set in advance in the facility, it may be used, and further, a message for guiding by voice may be sent.

  In the present embodiment, a camera is cited as an example of the second person flow measuring device 101b. However, in a facility having an existing surveillance camera or security camera, the surveillance camera or security camera may be used as the second person flow measuring device 101b. good.

As described above, according to the present embodiment, it is possible to set a human flow estimation target region even in an open space, and highly accurate human flow estimation based on measurement data from a plurality of human flow measurement devices having different measurement characteristics. It is possible to provide a human flow estimation device, a human flow estimation method, and a human flow estimation system.
Further, according to the present embodiment, it is possible to accurately estimate the human flow without requiring identification and tracking of individual pedestrians in an open space where no OD data acquisition device is installed.
Furthermore, a wide range of complex spaces that could not be covered by a single human flow measurement device can be combined with human flow measurement devices with different measurement characteristics to interpolate the situation where each measurement accuracy cannot be guaranteed, The estimation accuracy as the whole estimation apparatus is improved. Furthermore, by estimating the OD data by data assimilation, the position of the OD data to be estimated and the installation position of the human flow measurement device do not necessarily coincide with each other, and equipment such as existing surveillance cameras can be used. Operation costs can be reduced.

  FIG. 11 is an overall schematic configuration diagram of the human flow estimation system according to the second embodiment of the present invention, and FIG. 12 is a functional block diagram of the human flow estimation device constituting the human flow estimation system shown in FIG. is there. In the present embodiment, a completely open space on a street in the city is a target area for human flow estimation, and the human flow estimation device is connected to a portable terminal possessed by a human flow measurement device and a guide via a communication network, and This is different from the first embodiment. Constituent elements similar to those in the first embodiment are denoted by the same reference numerals, and redundant description of the first embodiment is omitted.

  As shown in FIG. 11, the human flow estimation system according to the present embodiment is possessed by the human flow estimation device 10a, two or more types of first human flow measurement devices 101a, second human flow measurement devices 101b having different measurement characteristics, guides, and the like. Mobile terminals 50a to 50c and a communication network 51 that connects these terminals so as to communicate with each other are provided. It does not matter whether the communication network 51 is wired or wireless. For convenience of explanation, only the mobile terminals 50a to 50c are shown, but it goes without saying that mobile terminals corresponding to the number of guides are provided. For example, a smartphone or a PDA (Personal Data Assistance) is used as the mobile terminals 50a to 50c.

  As shown in FIG. 12, the human flow estimation device 10a includes an input of an allowable measurement error based on specifications for each of the human flow measurement device 101a and the second human flow measurement device 101b, an input of an analysis model, and a virtual entrance described in detail later. An input unit 12, a processing unit 103, and a display unit 13 that can set a virtual exit are provided. The processing unit 103 includes an input I / F 104, a hypothesis OD generation unit 105, and a hypothesis OD generation unit 105 that capture an allowable measurement error, an analysis model, and a virtual entry and a virtual exit that are set from the input unit 12. A human flow analysis unit 106 that performs human flow analysis based on the generated OD data, a data assimilation OD estimation unit 108, a storage unit 109, a display control unit 110, and a communication I / F 111 are connected to each other via an internal bus 112. Has been. The hypothesis OD generation unit 105 and the human flow analysis unit 106 constitute an analysis unit 107. The hypothesis OD generation unit 105, the human flow analysis unit 106, the data assimilation OD estimation unit 108, and the display control unit 110 are, for example, a processor such as a CPU, a ROM that stores a program, and a program that is read from the ROM. This is realized by a storage device such as a RAM that temporarily stores process data and the like.

FIG. 13 is a diagram illustrating an analysis model in a case where a city which is a completely open space is a target area for estimating human flow. As shown in FIG. 13, the analysis model assumes that a human flow from a station 904 or another road toward the stadium 905 is estimated. On the road existing between the station 904 and the stadium 905, two kinds of first person flow measurement devices 101a and second person flow measurement devices 101b having different measurement characteristics are installed at a plurality of main positions. Each road is provided with a plurality of virtual doorways 801 a set by the user via the input unit 12. Here, it is desirable that the number of installed first person flow measurement apparatuses 101a and second person flow measurement apparatuses 101b is at least larger than the estimated number of OD data.
In step S11 of the flowchart of FIG. 2 shown in the first embodiment described above, as an initial setting, in addition to the permissible measurement error based on the specifications for each of the analysis model, the human flow measurement device 101a, and the second human flow measurement device 101b shown in FIG. The virtual entrance / exit 801a is input via the input unit 12, and is taken into the processing unit 103 by the input I / F 104. The captured analysis model, the permissible measurement error based on the specifications for each of the human flow measurement device 101a and the second human flow measurement device 101b, and the virtual entrance / exit 801a are input to the predetermined unit of the storage unit 109 via the internal bus 111 from the input I / F 104. Stored in the storage area. Other configurations and operations of the human flow estimation device 10a are the same as those in the first embodiment, and thus the description thereof is omitted.

FIG. 14 is a screen display example of a human flow estimation result displayed on the display unit 13 and the mobile terminals 50a to 50c constituting the human flow estimation device 10a shown in FIG. The human flow estimation device 10a operates in the same manner as in the first embodiment described above, and the human flow estimation result obtained by the data assimilation OD estimation unit 108 constituting the human flow estimation device 10a is displayed via the internal bus 111 and the display control unit 110. In addition to being displayed on the screen of the unit 13, it is also displayed on the screens of the portable terminals 50 a to 50 c possessed by the guide via the internal bus 111, the communication I / F 111, and the communication network 51. In FIG. 14, black circles represent pedestrians and white triangles represent guides 906. On the screen, the human flow guidance display 903 is indicated by a white arrow. The guide can easily grasp the degree of congestion on the road and how to guide the pedestrian by visual observation.
In FIG. 14, the congestion area is not displayed on the screen. However, as in the first embodiment, the congestion degree distribution may be represented by shading and a display form for specifying the congestion area may be used.

  As described above, according to the present embodiment, in addition to the effects of the first embodiment, since the human flow estimation result is displayed on the screen of the mobile terminal possessed by the guide, the guide is appropriate for an appropriate pedestrian. Can be easily guided. In addition, it is possible to change the placement of guides according to the congestion situation, and it is possible to efficiently guide the direction person, in other words, to guide the direction person with a small number of guides and to save labor. Become.

  In addition, this invention is not limited to an above-described Example, Various modifications are included. For example, the above-described embodiments have been described in detail for easy understanding of the present invention, and are not necessarily limited to those having all the configurations described. Further, a part of the configuration of one embodiment can be replaced with the configuration of another embodiment, and the configuration of another embodiment can be added to the configuration of one embodiment.

1, 1a ... human flow estimation system 10, 10a ... human flow estimation device 12 ... input unit 13 ... display unit 101a ... first human flow measurement device 101b ... second human flow measurement device 101n ..Nth person flow measuring device 102 ... measurement unit 103 ... processing unit 104 ... input I / F
105 ... Hypothesis OD generation unit 106 ... Human flow analysis unit 107 ... Analysis unit 108 ... Data assimilation OD estimation unit 109 ... Storage unit 110 ... Display control unit 111 ... Communication I / F
112 ... Internal bus 20 ... Congestion estimation target area 201 ... Entrance / exit A
202 ... Entrance / exit B
203 ... Entrance C
204 ... Measurement area 205 of the first human flow measurement device ... Measurement area 206 of the second human flow measurement device ... Structure 40 ... Advancing direction 401 ... Pedestrian 402 ... Upper left direction 403 ..Upward direction 404 ... Upper right direction 405 ... Right direction 406 ... Lower right direction 407 ... Down direction 408 ... Lower left direction 409 ... Left directions 50a, 50b, 50c ... Mobile Terminal 51 ... Communication network 60 ... Human flow analysis 601 ... Target pedestrian 602 ... Calculation grid 603 ... Entrance (O)
604 ... Exit (D)
605: Candidate route to exit 606 ... Shortest route 607 ... Selected route 608 ... Pedestrian 70 existing on shortest route ... Data assimilation 701 ... True value 702 ... Probability density distribution 703 of human flow analysis prediction value ... Probability density distribution 704 of measurement value ... Ensemble member 705 ... Post distribution 801 ... Entrance / exit 801a ... Virtual entrance / exit 802 ... Stairs 803 ...・ Elevator 90 ... Human flow estimation result display screen 901 ... Congestion degree distribution 902 ... Congestion area 903 ... Human flow guidance display 904 ... Station 905 ... Stadium 906 ... Guide

Claims (15)

The number of passing people and the direction of pedestrian travel obtained from two or more types of human flow measuring devices having different measurement characteristics are obtained, and the number of pedestrians heading from the actual entrance to the exit or the virtual entrance set is virtual. An analysis unit for generating a plurality of OD data, which is the number of pedestrians heading to the exit, as hypothetical OD data and performing a human flow analysis;
Data for generating estimated OD data by integrating the number of passing people and the pedestrian travel direction obtained from the human flow measurement device and the data obtained from the analysis unit based on an allowable measurement error determined based on the human flow measurement device An assimilation OD estimation unit;
A human flow estimation apparatus comprising: a display control unit that displays estimated OD data obtained from the data assimilation OD estimation unit on a display unit as a human flow estimation result. The human flow estimation device according to claim 1,
2. The human flow estimation device according to claim 1, wherein the pedestrian travel directions acquired from the human flow measurement device are two or more directions. The human flow estimation device according to claim 2,
The said display control part displays congestion degree distribution so that a congestion area | region can be identified to the human flow estimation result displayed on the said display part, The human flow estimation apparatus characterized by the above-mentioned. The human flow estimation device according to claim 3,
The display control unit displays a human flow guidance display capable of reducing congestion based on the congestion degree distribution on the display unit. The human flow estimation device according to any one of claims 2 to 4,
A human flow estimation apparatus, comprising: an input unit that inputs a target region to be subjected to the human flow analysis as an analysis model and allows the virtual entrance and the virtual exit to be set in the analysis model. Two or more types of human flow measuring devices having different measurement characteristics;
A human flow estimation device for performing human flow estimation based on at least the number of passing people and the direction of travel of the pedestrian measured by the human flow measurement device,
The human flow estimation device includes:
The number of pedestrians who go from the actual entrance to the virtual exit, or the number of pedestrians who go from the actual entrance to the exit by acquiring the passing number and the traveling direction of the pedestrian measured by the flow measuring device. An analysis unit that generates a plurality of OD data as hypothetical OD data and performs human flow analysis;
Data for generating estimated OD data by integrating the number of passing people and the pedestrian travel direction obtained from the human flow measurement device and the data obtained from the analysis unit based on an allowable measurement error determined based on the human flow measurement device An assimilation OD estimation unit;
A human flow estimation system comprising: a display control unit that displays estimated OD data obtained from the data assimilation OD estimation unit on a display unit as a human flow estimation result. The human flow estimation system according to claim 6,
2. The human flow estimation system according to claim 1, wherein the pedestrian travel directions acquired from the human flow measurement device are two or more directions. The human flow estimation system according to claim 7,
One type of human flow measurement device among the two or more types of human flow measurement devices having different measurement characteristics is capable of measuring the number of passing people and the direction of travel of pedestrians over a wide range. The human flow estimation system according to claim 7,
Of the two or more types of human flow measurement devices having different measurement characteristics, one type of human flow measurement device can measure the number of passing people and the direction of travel of pedestrians in a narrow range, and is installed even in an environment with many shields. A human flow estimation system that is possible. The human flow estimation system according to claim 8 or 9, wherein
The display control unit displays a congestion degree distribution so that a congestion region can be identified in a human flow estimation result displayed on the display unit. The human flow estimation system according to claim 10,
The said display control part displays the human flow guidance display which can relieve congestion based on the said congestion degree distribution on the said display part, The human flow estimation system characterized by the above-mentioned. The human flow estimation system according to claim 11,
A communication network that connects a plurality of portable terminals and the human flow measurement device and the human flow estimation device in a communicable manner;
The display control unit displays the estimated OD data on the plurality of portable terminals as a human flow estimation result via the communication network. The human flow estimation system according to claim 12,
The human flow estimation system includes an input unit that inputs a target region where the human flow analysis is to be performed as an analysis model and allows the virtual entrance and virtual exit to be set in the analysis model. . At least two types of human flow measurement devices having different measurement characteristics, and a human flow estimation method for performing human flow estimation based on the number of passing people and the direction of travel of pedestrians measured by the human flow measurement device,
The number of pedestrians who go from the actual entrance to the virtual exit, or the number of pedestrians who go from the actual entrance to the exit by acquiring the passing number and the traveling direction of the pedestrian measured by the flow measuring device. Generate multiple OD data as hypothetical OD data, perform human flow analysis,
Based on the allowable measurement error determined based on the human flow measurement device, the estimated number of passing people and the pedestrian travel direction obtained from the human flow measurement device and the data obtained from the result of the human flow analysis are integrated to generate estimated OD data. And
A human flow estimation method, wherein the generated estimated OD data is displayed on a display unit as a human flow estimation result. The human flow estimation method according to claim 14,
2. A human flow estimation method, wherein a pedestrian travel direction acquired from the human flow measurement device is two or more directions.

Images