JP2013173595A - Elevator arrival time estimating device and elevator system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a technique for accurately estimating a waiting time before an elevator car arrives at a floor.SOLUTION: An elevator arrival time estimating device includes: estimating the number of passengers by using an image of a camera provided in an elevator car; estimating the number of waiting passengers in a floor by using the image of the camera installed in a passenger platform of the floor; and estimating a time before the car arrives at the floor based on the estimated number of the passengers and of the waiting passengers and on operation control information of the elevator.

Description

  The present invention relates to a technique for controlling an elevator.

  In the conventional elevator control system, the user predicts the arrival time of the car by himself based on the current position of the car displayed on the door of the elevator hall. However, due to boarding / exiting on the middle floor, it may take more time than expected to feel stress. Therefore, a technique for predicting the waiting time of an elevator car has been developed.

  In the following Patent Document 1, the average number of people getting on and off on each floor is statistically calculated by using a scale device under the car and a camera in the car, thereby estimating the average number of people on and off the calling floor. The average waiting time is predicted based on this.

  Patent Document 2 below describes a technique for determining whether to stop / pass a car on each floor based on the video of a monitoring camera arranged on the stop floor of the elevator.

JP-A-8-91717 JP 2008-195468 A

  In the technique described in Patent Document 1, since the waiting number when a call is made is estimated based on past statistics, there is a possibility that the past average waiting number and the current waiting number are not necessarily equal. . Especially in a commercial facility such as a department store, the number of waiting persons greatly varies from 1 to 10 persons from time to time, so the average number of waiting persons when the call button is pressed in the past may be different from the current waiting number. Since this deviation becomes a large error with respect to the estimated arrival time, the waiting time cannot be predicted correctly.

  In the technique described in Patent Document 2 described above, an attempt is made to optimize whether each car stops or passes on each floor in order to reduce the waiting time. The point presented to the person is insufficient.

  The present invention has been made to solve the above-described problems, and an object of the present invention is to provide a technique for accurately estimating the waiting time until an elevator car reaches the floor.

  The elevator arrival time estimation device according to the present invention estimates the number of passengers using the images of the cameras provided in the elevator car, and estimates the number of people waiting on the floor using the images of the cameras installed at the boarding / alighting locations on the floor. Then, based on the estimated number of passengers and waiting persons and the elevator operation control information, the time until the car arrives on the floor is estimated.

  According to the elevator arrival time estimation device according to the present invention, it is possible to accurately estimate the arrival time of each level of the elevator. Further, by displaying the estimation result as the remaining waiting time on each floor, it is possible to ease the burden on the customer or to prompt the user to select another lifting / lowering means, thereby improving user convenience.

1 is a configuration diagram of an elevator system 1000 according to Embodiment 1. FIG. 3 is a functional block diagram illustrating a configuration of an arrival time estimation unit 100. FIG. It is a figure explaining the method for the passenger number estimation part 120 to estimate the number of passengers in the car. It is a figure explaining the other method in which the passenger number estimation part 120 estimates the passenger number in the cage | basket | car 400. FIG. It is a flowchart of the procedure demonstrated in FIG. Hereinafter, each step of FIG. 5 will be described. It is a figure which shows typically the result example which the waiting number estimation part 130 estimated the number of people waiting for the car 400 in the elevator hall of each floor. It is a figure which shows the example of the present operation control information of the elevator which the elevator information acquisition part 110 acquires from the elevator control part 200. FIG. It is a figure which shows the structure of past example DB150. It is a figure which shows the structural example of car control DB152. It is a figure which shows the structural example of passenger DB153. It is a figure which shows the structural example of floor information DB154. It is a figure which shows the structural example of waiting customer DB155. It is a flowchart explaining the procedure in which the arrival time calculation part 140 calculates the arrival time of the car 400 using the past case DB150. It is a functional block diagram which shows the structural example of the arrival time estimation apparatus 1400 which mounted the arrival time estimation part 100 with the computer. 3 is a functional block diagram of an elevator control unit 200. FIG. It is a figure which shows the example in which the car call terminal 601 (or 602) displays the waiting time until the car 400 arrives. It is a processing flow in which the elevator control unit 200 controls opening and closing of the car side door 430 and the hall side doors (611, 612) according to the above method. This is a processing flow in which the overall control unit 240 of the elevator control device 200 controls the moving speed of the car 400. FIG. 5 is a schematic diagram showing a state where a car A and a car B exist, the destination floors specified on each car operation terminal 410 are different, and the car is called on a plurality of floors. In Embodiment 5, it is a figure explaining the method in which the boarding number estimation part 120 and the waiting number estimation part 130 detect a person.

<Embodiment 1>
FIG. 1 is a configuration diagram of an elevator system 1000 according to Embodiment 1 of the present invention. The elevator system 1000 includes an arrival time estimation unit 100, an elevator control unit 200, cameras outside the car 301 and 302, an in-car camera 410, and car calling terminals 601 and 602.

  The outside-car cameras 301 and 302 are installed in an elevator hall (elevator boarding place) on the floor where the elevator stops. The in-car camera 410 is installed in the elevator car 400. The outside-car cameras 301 and 302 and the in-car camera 410 are connected to the arrival time estimation unit 100 via the communication infrastructure 500. Video captured by each camera is transmitted to the arrival time estimation unit 100 via the communication infrastructure 500. The communication infrastructure 500 is a communication network such as a LAN (Local Area Network) or a video transmission cable.

  The car call terminals 601 and 602 are operation terminals for calling the car 400, and are installed on each floor where the car 400 stops. The car calling terminals 601 and 602 can also open and close the hall side doors 611 and 612, respectively. In addition, the car calling terminals 601 and 602 display estimated arrival times described later.

  A car operation terminal 420 is installed in the car 400. The car operation terminal 420 is an operation terminal for specifying a floor on which the car 400 is moved and for opening and closing the car side door 430.

  The arrival time estimation unit 100 estimates the number of passengers in the car 400 and the number of waiting passengers waiting for an elevator on each floor based on images taken by each camera by image recognition. The arrival time estimation unit 100 estimates the estimated time for the car 400 to arrive at each floor based on the number of persons and the elevator operation control information obtained from the elevator control unit 200, and outputs the estimated time to the elevator control unit 200.

  The elevator control unit 200 is a device that controls the operation of the elevator. The elevator control unit 200 transmits the estimated arrival time of the car 400 to the car call terminals 601 and 602 and the car operation terminal 420. The car calling terminals 601 and 602 display the received estimated arrival time in a display format as shown in the lower right of FIG. As a result, a waiting customer waiting for the car 400 on each floor can know the waiting time until the calling car 400 arrives, so that the waiting stress can be reduced, and stairs, escalators, etc. The estimated arrival time can be used as information for determining whether to select an alternative means.

  Cameras 301 and 302 outside the car are preferably installed on all floors where the car 400 stops, but may be installed only on specific floors such as the first floor and the third floor. A plurality of cameras outside the car may be installed on each floor. By installing multiple cameras outside the car on a very busy floor such as the hall with a large hall or the first floor, all waiting customers can be photographed with the camera, so the estimated arrival time can be estimated accurately. Can do.

  FIG. 2 is a functional block diagram showing the configuration of the arrival time estimation unit 100. As shown in FIG. The arrival time estimation unit 100 includes an elevator information acquisition unit 110, a passenger number estimation unit 120, a waiting number estimation unit 130, an arrival time calculation unit 140, and a past case database (DB) 150.

  The elevator information acquisition unit 110 acquires elevator operation control information from the elevator control unit 200. The operation control information is information indicating the operation status such as whether the elevator is currently operating or stopped, on which floor it is scheduled to stop, from which floor the calling instruction is generated, and how much kg the total weight in the car is. is there. Details will be described later.

  The number-of-passengers estimation unit 120 acquires the images captured by the car camera 410 and estimates the number of passengers in the car 400 based on the images. The waiting number estimating unit 130 acquires images taken by the cameras 301 and 302 outside the car, and estimates the number of people who have the car 400 in the elevator hall on each floor based on this. These estimation results are output to the arrival time calculation unit 140 and recorded in the past case DB 150.

  The arrival time calculation unit 140 calculates the estimated arrival time of the car 400 on each floor based on the operation control information acquired by each of the function units, the number of passengers, and the number of waiting persons, and transmits it to the elevator control unit 200.

  FIG. 3 is a diagram for explaining a method in which the passenger number estimation unit 120 estimates the number of passengers in the car 400. It is assumed that the car camera 410 has photographed the passengers in the car 400 as shown in FIG. The passenger number estimation unit 120 can detect a person in the car 400 by using, for example, a person detection technique described in Reference Document 1 below on the image illustrated in FIG. .

FIG. 3B shows an example of a HOG (Histogram of Oriented Gradient) feature quantity used for human detection. When calculating the feature amount, a person's head is divided into a plurality of small regions, HOG feature amounts indicating the gradient direction and strength of the edges in the small region are extracted, SVM (Support Vector Machine), AdaBoost, etc. By using the classifier constructed using the machine learning technique, it is discriminated whether the region in the image is a human head or the other. Thus, a person can be detected from the video, and a detection result as shown in FIG. 3C can be obtained.
[Reference 1] Chengbin Zeng, Huadong Ma, “Robust Head-shoulder Detection by PCA-Based Multilevel HOG-LBP Detector for People Counting”, 2010 20th International Conference on Pattern Recognition (2010), pp. 2069-2072.

  In addition to the HOG feature value, another feature value such as Haar or LBP (Local Binary Pattern) may be used. By using this method, it is possible to detect only the area of the person's head, so that each person can be detected and the correct number of people can be estimated even in situations where people are dense and the body of the person behind is not visible. .

  FIG. 4 is a diagram illustrating another method in which the passenger number estimation unit 120 estimates the number of passengers in the car 400. The method described in FIG. 3 is resistant to overlapping human bodies, but there is a problem that if the head is hidden, the person cannot be detected and the number of persons cannot be estimated correctly. Therefore, the present invention proposes a method for detecting a person whose head is partially hidden by supplementarily using foreground extraction based on background difference. Hereinafter, this method will be described with reference to FIG.

  It is assumed that the car camera 410 has photographed the passengers in the car 400 as shown in FIG. The number-of-passengers estimation unit 120 extracts a foreground region by taking a difference between a background image and a current image in a state in which a prepared person is not on the vehicle. FIG. 4B is a diagram in which the foreground area is filled. The number-of-passengers estimation unit 120 detects a person existing in the foreground portion by performing the head detection described with reference to FIG. 3 on the detected foreground region (FIG. 4C). The number-of-passengers estimation unit 120 estimates the number of persons present in the portion of the foreground area where the head is not detected based on the area (FIG. 4D). Details of this method will be described with reference to FIG.

  FIG. 5 is a flowchart of the procedure described in FIG. Hereinafter, each step of FIG. 5 will be described.

(FIG. 5: Steps S501 to S503)
The number-of-passengers estimation unit 120 performs each procedure described with reference to FIGS. 4 (a) to 4 (c).

(FIG. 5: Step S504)
The number-of-passengers estimation unit 120 estimates the attributes of the person detected in step S503. For example, (a) a type of luggage such as a stroller, a wheelbarrow, a cane, a trunk, (b) an age type such as an elderly person or an infant, and (c) a person type such as a customer or an employee are estimated as person attributes. Can be considered. For example, if pattern matching is performed with a typical image of an assumed person attribute and the similarity between both images is equal to or greater than a predetermined threshold, it can be estimated that the person has the attribute.

(FIG. 5: Step S504: Supplement)
The attribute information calculated in this step can be used to correct the time required for the detected person to get on and off the car 400. For example, since it seems that it takes time to get on and off a person with a lot of luggage or an older person, it is possible to correct the boarding time by multiplying the standard boarding time of other persons by a coefficient. Since this step is a preliminary step for improving the estimation accuracy, it is not necessarily performed.

(FIG. 5: Step S505)
The number-of-passengers estimation unit 120 removes a region considered to be a person region based on the head region detected in step S503 from the foreground region detected in step S502, and calculates the position and area of the remaining foreground region. This step corresponds to the process of calculating the filled area shown in FIG.

(FIG. 5: Step S506)
The number-of-passengers estimation unit 120 estimates the number of persons estimated that the head was not detected but would exist in the foreground area based on the area and location of the remaining foreground area calculated in step S505. . For example, for each foreground area i, the area of the foreground area i is Ai, the area of the person assumed when a person is present at the position of the foreground area i is Si, and Ai ÷ Si is greater than or equal to a predetermined threshold value If there is one person in that part, it is estimated. By performing this process for all foreground areas I (iεI), the number of persons existing in the remaining foreground area can be estimated.

(FIG. 5: Step S506: Supplement)
The area Si of the person used in this step needs to be changed according to the position of the foreground region i. For example, Si is increased at a position close to the in-car camera 410, and Si is decreased at a far position. The number of passengers obtained in this step may be an accurate number of passengers or may be used as a value indicating a more ambiguous congestion degree.

  FIG. 6 is a diagram schematically illustrating an example of a result of estimation of the number of people waiting for the car 400 in the elevator hall on each floor by the waiting number estimating unit 130. The waiting number estimating unit 130 detects a person's head by the same method as the number of passengers estimating unit 130 and estimates the number of people in the elevator hall. However, since the installation position and the depression angle of the camera outside the car are different from those of the in-car camera 410, the resolution of the person existing in the video is also different from that of the car camera 410. Therefore, parameters such as the minimum and maximum size of the head area to be detected need to be adjusted for the camera outside the car.

  When a plurality of elevators are operating and there are a plurality of hall-side doors (door A611a and door B611b in the example shown in FIG. 6) in the elevator hall, which elevator each standby customer is waiting for Need to be identified. In this case, the waiting number estimating unit 130 estimates which elevator is waiting for each head region based on the detected head position and door position. For example, it is estimated that the elevator is waiting for the door having the closest distance between the center of gravity of the head and the center of gravity of the door. Thereby, as shown in FIG. 6, the waiting number of persons for every elevator car can be estimated.

  FIG. 7 is a diagram illustrating an example of the current operation control information of the elevator that the elevator information acquisition unit 110 acquires from the elevator control unit 200. The operation control information includes an overall control information table 111 and a car information table 112. In each table, a record is added every time the elevator control unit 200 issues or acquires control information. Here, a data format in which a column is added in the right direction every time a record is added is illustrated. The same applies to the database described below.

  The overall control information table 111 holds information related to the operation status of the entire elevator system 1000. Specifically, information about the current operation status of each car, the floor on which the up call is made, the floor on which the down call is made, and the like are held. The elevator control unit 200 can acquire such information every time a call operation is performed on the car call terminal on each floor. By keeping the time when the call button on each floor is pressed, it is possible to know how long the waiting time has elapsed on each floor.

  The car information control table 112 holds information regarding the operation status of a specific car. The car information control table 112 is provided for each car. FIG. 7 illustrates the car information control table 112 for the car A. Each time the car 400 closes the car-side door 430, the elevator control unit 200 is called the time at that time, the ascending / descending classification, the currently stopped floor, and the target floor specified by the car operating terminal 420. And the weight in the car 400 obtained from the scale installed under the car, the presence / absence of boarding / exiting before and after closing the car-side door 430, and the like as records on the car information control table 112 to add.

  FIG. 8 is a diagram showing the configuration of the past case DB 150. As shown in FIG. The past case DB 150 holds an overall control DB 151, a car control DB 152, a passenger DB 153, a floor information DB 154, and a standby customer DB 155 as sub-DBs.

  The overall control DB 151 is a DB that stores past cases regarding the operation control information of the entire elevator system 1000, and accumulates the overall control information table 111 obtained by the elevator information acquisition unit 110 in a time series as it is. Thereby, it is possible to grasp a change in operation, a situation where the call button is pressed, and the like.

  FIG. 9 is a diagram illustrating a configuration example of the car control DB 152. The car control DB 152 is a DB in which operation control information related to a car is compiled. When there are a plurality of cars, a different DB is created for each car. The information stored in the car control DB 152 is obtained by adding the number of passengers of each car estimated by the passenger number estimating unit 120 to the contents of the car information table 112 acquired by the elevator information acquiring unit 110. Thereby, each car can grasp when and where how many passengers were carried. Each record may be stored each time the elevator information acquisition unit 110 acquires the car information table 112, for example.

  FIG. 10 is a diagram illustrating a configuration example of the passenger DB 153. The passenger DB 153 is provided for each car corresponding to each car control DB 152. For example, the passenger DB 153 of the car A is provided corresponding to the car control DB 152 of the car A. Passenger DB153 hold | maintains the information about the passenger which the passenger number estimation part 120 recognized. Specifically, the passenger ID, the boarding floor, the boarding time, the boarding floor, the boarding time, and attributes are stored. In a congested environment such as an elevator, it is difficult to specify the getting-off floor by continuously tracking passengers. If the passenger cannot be specified, “**” or the like is recorded for the passenger ID. Each record may be stored, for example, every time the number of passengers estimation unit 120 estimates the number of passengers.

  FIG. 11 is a diagram illustrating a configuration example of the floor information DB 154. The floor information DB 154 is a DB that accumulates operation control information related to each floor, and an individual DB is created for each floor. The elevator information acquisition unit 120 adds a record to the floor information DB 154 every time the car call status of each floor changes. The waiting number estimating unit 130 adds a record to the floor information DB 154 every time the estimation result of the waiting number on each floor changes. With this DB, it is possible to grasp when a car call occurs on each floor, how many people are waiting on the floor, and whether the waiting person is on the car.

  FIG. 12 is a diagram illustrating a configuration example of the standby customer DB 155. The waiting customer DB 155 is a DB that accumulates logs about waiting customers on each floor, and an individual DB is created for each floor. The waiting number estimating unit 130 adds a record to the waiting customer DB 155 every time the number of waiting persons on each floor is estimated. The standby customer DB 155 stores a standby start time, a standby position, a boarding time, a boarded car, attributes, and the like based on the estimation result by the waiting person estimation unit 130. Thereby, it is possible to manage what kind of customers have been waiting on each floor.

  FIG. 13 is a flowchart for explaining the procedure by which the arrival time calculation unit 140 calculates the arrival time of the car 400 using the past case DB 150. This flowchart is executed every time the arrival time calculation unit 140 calculates the estimated arrival time on each floor. Hereinafter, each step shown in FIG. 13 will be described.

(FIG. 13: Step S1301)
The arrival time calculation unit 140 acquires, as current elevator operation control information, the calling status of the car 400 on each floor, the number of people waiting on each floor, and the number of passengers in each car 400. The calling status of the car 400 and the number of waiting persons can be acquired from the floor information DB 154. The number of passengers in each car 400 can be acquired from the car control DB 152.

(FIG. 13: Step S1302)
The arrival time calculation unit 140 compares each value acquired in step S1301 as a multi-dimensional feature quantity with a further previous record stored in the past case DB 150, thereby waiting on each floor, the number of passengers, and the time. Search for records with similar bands. The inter-case distance D (x, y) between the current state x and the past case y at this time can be expressed by the following equation 1, for example.

(FIG. 13: Step S1302: Formula)

  In Equation 1, M represents the number of passengers, and Wf represents the number of people on the floor f. That is, Mx represents the number of passengers in the car in the current state x, and Wfx represents the current number of people waiting on the floor f. Fg represents the destination floor of the car 400, and Fn represents the floor where the car 400 is currently located. Formula 1 is a function that returns a smaller value as the current number of passengers / standby number of persons approaches the past number of passengers / number of standby persons. According to Equation 1, the inter-event distance D between the current situation and the past case can be obtained.

(FIG. 13: Step S1302: Supplement 1)
Depending on the usage environment of the elevator system 1000, the arrival time may be completely different depending on the time zone even if the number of passengers and the waiting number are the same. In this case, the range in which the current operation control status is compared with the past case is changed to the current status. You may limit only to the near range. For example, the past case DB 150 may be searched only for the past history of the same day of the week and the same time zone as the current date and time.

(FIG. 13: Step S1302: Supplement 2)
In this step, if the attributes of waiting passengers on each floor and passengers in the car 400 are known, the number of passengers and the number of waiting persons for each attribute are set as variables in Equation 1, and the distance between cases is obtained for each attribute. May be added together. Furthermore, when the boarding / alighting time varies depending on the attribute, a variable representing the number of passengers and the number of waiting persons for each attribute may be multiplied by a coefficient representing the boarding / alighting time of the attribute. For example, since a person with a large load takes time to get on and off, it is conceivable to calculate by multiplying by a factor of 1.5 and for 1.5 persons.

(FIG. 13: Step S1303)
The arrival time calculation unit 140 extracts the top K past cases having a small inter-case distance D from the current situation from the past case DB 150, and calculates the average distance between them. If the average distance is smaller than a threshold value T, it is determined that there is a certain past case, and the process proceeds to step S1304. If the average distance is equal to or greater than the threshold value T, it is determined that there is no certain past case, and the process proceeds to step S1305.

(FIG. 13: Step S1304)
The arrival time calculation unit 140 extracts the time required to arrive at each floor for the K past cases extracted in step S1303, and the average or median value of the K cases is obtained as the arrival time estimation result. To do. However, if there is a large variation in the time required for arrival in the K past cases, it is considered that the reliability of the case is low, and the process proceeds to step S1305 without adopting the K cases.

(FIG. 13: Step S1305)
When this step is reached, there is no appropriate similar case in the past case DB 150, so it is necessary to estimate the estimated arrival time without using the past case. Therefore, the arrival time calculation unit 140 calculates the required time Tng from the current car position Fn to the specific floor Fg based on the following equation 2, for example. Tng can be calculated using the car stop schedule Sfε {0, 1} at each floor f (Fn <f <Fg), the waiting number Wf of each floor f, and the current number of passengers M.

(FIG. 13: Step S1305: Formula)

  In Equation 2, V is the average travel time, and α, β, and γ are fixed parameters. The first term of Equation 2 is the time required when the current car position to the specific floor Fg has never stopped. Since Sf in the second term is a function that returns 1 for the planned stop floor f and 0 otherwise, the required time Tng increases every time a stop floor occurs. α indicates the time required for getting on and off per waiting passenger, β indicates the fixed stop time of the elevator generated for each stop floor, and γ indicates the elevator stop time generated for each current passenger.

(FIG. 13: Step S1305: Supplement 1)
The actual arrival time on each floor can be obtained from the elevator information acquisition unit 110 while operating the elevator. By correcting the parameters α, β, and γ using the actual arrival time and the estimated arrival time calculated using Equation 2, the accuracy of Equation 2 can be improved. Although the value of this parameter varies greatly depending on the size of the car 400, the number of floors, and the introduction environment (type of office, store, etc.), the above correction eliminates the need for manual parameter optimization according to the environment.

(FIG. 13: Step S1305: Supplement 2)
About the value of Wf and M in Formula 2, you may multiply a coefficient according to the attribute of a waiting customer or a passenger like the matter explained at Step S1302. Thereby, the precision of Formula 2 can be raised. It can also be used in combination with correction relating to the parameters α, β, and γ.

(FIG. 13: Step S1306)
The arrival time calculation unit 140 determines the arrival time of the car 400 for each floor based on the required time to arrival obtained in step S1304 or S1305.

  FIG. 14 is a functional block diagram illustrating a configuration example of an arrival time estimation apparatus 1400 in which the arrival time estimation unit 100 is implemented by a computer. The arrival time estimation apparatus 1400 includes an I / F 1401, an image memory 1402, a CPU (Central Processing Unit) 1403, a RAM (Random Access Memory) 1404, a ROM (Read Only Memory) 1405, and a storage unit 1406.

  The I / F 1401 transmits / receives data to / from an external device. The elevator information acquisition unit 110, the passenger number estimation unit 120, the waiting number estimation unit 130, and the arrival time calculation unit 140 transmit and receive each data via the I / F 1401. The image memory 1402 is a storage device that stores video received from each camera.

  The CPU 1403 executes the functions stored in the elevator information acquisition unit 110, the passenger number estimation unit 120, the waiting number estimation unit 130, and the arrival time calculation unit 140 by executing a program stored in the ROM 1405. The RAM 1404 stores data that is temporarily used when the CPU 1403 operates. The ROM 1405 stores a program that realizes the functions of the functional units included in the arrival time estimation unit 100. The storage unit 1406 is configured using a storage device such as an HDD (Hard Disk Drive) and stores data held by the past case DB 150.

  FIG. 15 is a functional block diagram of the elevator control unit 200. The elevator control unit 200 includes a vehicle allocation control unit 210, an input unit 220, a display control unit 230, and an overall control unit 240.

  The vehicle allocation control unit 210 performs control such as elevator elevation, speed adjustment, door opening and closing. The input unit 220 acquires operation information on the car call terminal on each floor and operation information on the car operation terminal 420 in the car 400. The display control unit 230 receives the estimated arrival time of the car 400 for each floor obtained from the arrival time estimation unit 100 and outputs it to the car calling terminal of each floor. The overall control unit 240 controls the operation of each function unit and serves as an interface with the arrival time estimation unit 100.

  FIG. 16 is a diagram illustrating an example in which the car calling terminal 601 (or 602) displays a waiting time until the car 400 arrives. The car call terminal 601 calculates the remaining time until arrival based on the arrival time estimated by the arrival time estimation unit 100 and displays it on the display on the car call terminal 601.

  FIG. 16A shows an example in which the remaining waiting time is displayed on a scale. A scale is displayed on the display section on the side where the button is pressed, and the time until arrival can be grasped by the number of lighting of the scale. FIG. 16B is an example in which the waiting time is displayed as a numerical value. Detailed time can be grasped. Further, by displaying the number of waiting persons on each floor, auxiliary information for the user himself / herself to estimate the arrival time can be presented. Note that the display device shown in FIG. 16 may be installed in another place as a single display device instead of on the car call terminal 601. A similar display device may be arranged on the car operation terminal 420 in the car 400.

<Embodiment 1: Summary>
As described above, the elevator system 1000 according to the first embodiment uses the in-car camera 410 and the out-of-car cameras 301 and 302 to estimate the number of passengers in the car 400 and the number of waiting persons on each floor, and based on this. The time until the car 400 arrives on each floor can be estimated. Further, by displaying the estimated arrival time on each floor, it is possible to reduce the waiting time burden of the elevator user.

  In addition, the elevator system 1000 according to the first embodiment searches the past case DB 150 for past cases similar to the number of passengers in the car 400 and the number of waiting persons on each floor, and stores each case based on the waiting time in the past case. The time until the car 400 arrives is estimated. Thereby, an accurate estimation result based on an actual case can be obtained.

<Embodiment 2>
In the second embodiment of the present invention, the estimation result of the waiting number on each floor and the estimation result of the number of passengers in the car 400 described in the first embodiment are used, and the car side door 430 and the hall side doors (611, 612) are changed. An operation example for efficiently performing opening / closing control will be described. Since the configuration of the elevator system 1000 is the same as that of the first embodiment, the following description will focus on the operation relating to door opening / closing control.

  When the car 400 arrives at the floor and the door is opened, if the waiting customer has already got on the other car 400 or if all the waiting customers got on the car 400, the door closing button of the car operation terminal 420 If the user has forgotten to press, the car 400 stops uselessly for several seconds. In the second embodiment, in order to reduce the useless stop time and the elevator arrival time, the waiting number obtained by the waiting number estimating unit 130 and the method used by the riding number estimating unit 120 described with reference to FIG. 4 are used. The door is automatically opened and closed based on the movement information in the car 400 obtained by obtaining the difference between the background images when estimating the number of passengers.

  FIG. 17 is a processing flow in which the elevator control unit 200 controls the opening and closing of the car side door 430 and the hall side doors (611, 612) according to the above method. This processing flow is started when a predetermined time (for example, several seconds) elapses without the door opening button being pressed after the elevator control unit 200 opens these doors. Hereinafter, each step of FIG. 17 will be described.

(FIG. 17: Step S1701)
The elevator control unit 200 makes an inquiry to the waiting number estimating unit 130 to check whether or not the waiting number on the floor where the car 400 is currently stopped is zero. If the number of waiting persons is zero, it is determined that there is no passenger getting on and off the car 400, and the process proceeds to step S1703. If not, the process proceeds to step S1702.

(FIG. 17: Step S1702)
The elevator control unit 200 checks, for example, for several seconds, whether there is a passenger movement in the elevator hall on the floor where the car 400 is currently stopped. If there is no movement, it is determined that there is no passenger getting on and off the car 400, and the process proceeds to step S1703. If there is a movement, the process proceeds to step S1704.

(FIG. 17: Step S1702: Supplement)
The movement information in the elevator hall on the floor can be obtained by obtaining the difference between the background images when the waiting number estimating unit 130 estimates the waiting number in the floor using the method described in FIG.

(FIG. 17: Steps S1701 to S1702: Supplement)
In these steps, before moving to step S1703, the number of passengers estimation unit 120 obtains the difference between the background images when estimating the number of passengers using the method described in FIG. Based on the movement information, it may be determined whether there is a passenger who is going to get off the car 400. When there is a passenger who is going to get off the car 400, the process does not proceed to step S1703 but proceeds to step S1702 / S1704, respectively.

(FIG. 17: Step S1703)
The elevator control unit 200 determines that there is no passenger getting on and off the car 400 and closes the car door 430 after closing the hall doors (611, 612).

(FIG. 17: Step S1704)
Since the elevator controller 200 holds the door open in the current loop, the elevator controller 200 returns to step S1701 and determines again whether or not the door should be automatically closed.

<Embodiment 2: Summary>
As described above, the elevator system 1000 according to the second embodiment automatically uses the estimation results of the waiting number estimating unit 130 and the passenger number estimating unit 120 to automatically lift the elevator door when the boarding / exiting on each floor is completed. And the car 400 can be started. As a result, it is possible to perform an efficient operation that saves useless waiting time. In addition, even when it is difficult for the passenger to carry the luggage or the like and operate the operation terminal, the car 400 can be started by automatically determining completion of boarding / exiting.

<Embodiment 3>
In the third embodiment of the present invention, an operation example in which the waiting time is shortened by increasing the moving speed of the car 400 under the configuration described in the first and second embodiments will be described. The car 400 may be configured to increase the moving speed according to the number of passengers. In the third embodiment, an operation example will be described under the assumption.

  FIG. 18 is a processing flow in which the overall control unit 240 of the elevator control device 200 controls the moving speed of the car 400. Hereinafter, each step of FIG. 18 will be described.

  The overall control unit 240 checks whether or not the moving speed of the car 400 can be increased based on the current number of passengers of each car obtained from the arrival time estimation unit 100 (S1801). For example, based on the measurement result of the weight measuring unit attached to the car 400, it may be determined that the speed can be increased if the loaded weight is less than a predetermined value. If the speed can be increased, the process proceeds to step S1802, and if not, the process proceeds to step S1804.

  The overall control unit 240 determines whether or not the waiting number of persons on each floor estimated by the waiting number estimating unit 130 is equal to or greater than a predetermined threshold. If it is equal to or greater than the threshold, the process proceeds to step S1803, and otherwise, the process proceeds to step S1804 (S1802).

  If it is determined that the moving speed of the car 400 can be increased, the overall supervision unit 240 issues a control command to the dispatch control unit 210 to increase the lifting speed of the car 400 (S1803). Otherwise, nothing is done because there is no speed change (S1804).

<Embodiment 3: Summary>
As described above, the elevator system 1000 according to the third embodiment can be operated by increasing the ascending / descending speed of the car 400 when there are many waiting persons on each floor. Further, when it is quiet, even when the lifting speed can be increased, the car 400 can be operated at the normal speed, and unnecessary energy consumption can be suppressed.

<Embodiment 4>
In the fourth embodiment of the present invention, an operation example of optimizing a vehicle allocation plan in an elevator system 1000 that controls a plurality of cars 400 will be described. Since the configuration of the elevator system 1000 is the same as that of the first to third embodiments, the following description will focus on the differences regarding the vehicle allocation plan for the plurality of cars 400.

  FIG. 19 is a schematic diagram showing a state in which the car A and the car B exist, the destination floors specified on the car operation terminals 410 are different, and the car is called on a plurality of floors. Based on the estimated arrival time of the car 400 for each floor estimated by the arrival time estimating unit 100, the current number of people waiting on each floor, and the number of passengers in the car 400, the overall supervision unit 240 formulates a car allocation plan for each car. Then, the vehicle allocation control unit 210 is executed. The allocation plan is based on the following three rules.

(Vehicle allocation plan rule 1)
A vehicle allocation plan is made so that the maximum waiting time among the waiting times on each floor estimated by the arrival time estimation unit 100 is as small as possible. That is, a vehicle allocation plan is created according to the so-called minimax method.

(Vehicle allocation plan rule 2)
Even if the car is called on the floor where the car is moving, it is not the destination floor of the car but the destination floor of the other car, and the waiting number on the floor where the car is called can be reliably accommodated for other cars If you have a small number of people, do not stop,

(Vehicle allocation plan rule 3)
If a car is called on the floor where the car is moving and the floor on which the car is called is not the target floor of any car, the rule is that if any car stops on the floor on which the car is called According to whether 1 can be satisfied, the car to be stopped on the floor where the call is made is determined.

  In FIG. 19, the car A is on the 19th floor, the car B is on the 18th floor, the destination floors designated on the car operation terminal 420 of the car A are the 17, 16, and 1st floors, and the car operation of the car B The destination floor designated on the terminal 420 is the first floor, and the car is called on the 17th floor, the 15th floor, and the first floor. In this state, it is considered that the rules 1 to 3 are applied.

  Regarding Rule 1, it is assumed that the largest of the waiting times on each floor is the waiting time on 1F which is far from any car. According to rule 1, car B will be directed to 1F.

  On the way of the car B toward 1F, the car is called 17F and 15F before the car A, but according to the rule 2, only the car A is controlled to stop at 17F. Even if the car B is lower than the car A, the waiting time on the first floor is longer than the waiting time on the 15th floor, so the car B does not stop on the 15th floor according to the rule 3.

<Embodiment 4: Summary>
As described above, the elevator system 1000 according to the fourth embodiment optimally manages a plurality of cars according to the rules 1 to 3 described above, and keeps the maximum waiting time as short as possible. be able to.

<Embodiment 5>
In the fifth embodiment of the present invention, a method for detecting a person at high speed in an image captured by each camera will be described. Since the configuration of the elevator system 1000 is the same as that of the first to fourth embodiments, the following description will focus on the differences related to the method for detecting a person.

  FIG. 20 is a diagram illustrating a method in which the number of passengers estimation unit 120 and the standby number estimation unit 130 detect people in the fifth embodiment. In order to detect a person in the image, it is necessary to search whether or not there is a part close to the shape of the person in the image. Specifically, pattern matching is performed with a human shape prepared in advance. Since this is performed on the entire image to search for the shape of a person, the process of detecting a person requires a long processing time.

  In pattern matching, the target image is divided into a plurality of window areas, the feature amount is calculated for each window, and compared with the feature amount of the comparison target image (human image if human detection). Make a decision. The feature value f (x1 + x, y1 + y) calculated at a certain position (x1 + x, y1 + y) in a certain image window (x1, y1, w, h) is generally used only for the matching determination of the image window. is there. For example, different feature quantities f (x1 + step_x + x, y1 + y) are calculated for the adjacent image window (x1 + step_x, y1, w, h) that is separated by step_x pixels in the x direction.

  In the fifth embodiment, in order to solve the above-described problem, a feature value f is calculated for each step_x and step_y that is an interval between image windows, and is stored in a memory such as a RAM. To reduce the amount of calculation.

  The small square in FIG. 20A shows an example of the HoG feature amount in a certain direction. As shown in FIG. 20B, an identifier that determines whether a certain image window is an object to be detected is also prepared such that it is regularly arranged for each step_x and step_y. Thereby, between the adjacent image window 1 and image window 2 shown in FIG. 20B, six feature amounts among the nine feature amounts can be shared. Since these feature amounts are stored in the memory, it is not necessary to calculate the feature amounts again.

<Embodiment 5: Summary>
As described above, the elevator system 1000 according to the fifth embodiment divides video captured by the camera into image windows, calculates feature amounts for each image window, and shares feature amounts between adjacent image windows. For example, if the number of feature quantities in the discriminator is F, the number of image windows in the x direction is W, and the number of image windows in the y direction is H, conventionally, the number of times that the feature quantity is calculated is F × W × H times. However, in the fifth embodiment, it can be greatly reduced to W × H times. Accordingly, the candidate positions of the detection target object can be narrowed down at high speed and detailed identification can be performed, so that high-speed and accurate object detection can be performed.

  The present invention is not limited to the embodiments described above, and includes various modifications. The above embodiment has been described in detail for easy understanding of the present invention, and is not necessarily limited to the one having all the configurations described. A part of the configuration of one embodiment can be replaced with the configuration of another embodiment. The configuration of another embodiment can be added to the configuration of a certain embodiment. Further, with respect to a part of the configuration of each embodiment, another configuration can be added, deleted, or replaced.

  Each of the above-described configurations, functions, processing units, processing means, and the like may be realized in hardware by designing a part or all of them, for example, with an integrated circuit. Each of the above-described configurations, functions, and the like may be realized by software by interpreting and executing a program that realizes each function by the processor. Information such as programs, tables, and files for realizing each function can be stored in a recording device such as a memory, a hard disk, an SSD (Solid State Drive), or a recording medium such as an IC card, an SD card, or a DVD.

  100: Arrival time estimation unit, 110: Elevator information acquisition unit, 111: Overall control information table, 112: Car information table, 120: Passenger number estimation unit, 130: Waiting number estimation unit, 140: Arrival time calculation unit, 150: Past case database, 151: Overall control DB, 152: Car control DB, 153: Passenger DB, 154: Floor information DB, 155: Standby customer DB, 200: Elevator control unit, 210: Vehicle allocation control unit, 220: Input unit, 230: Display control unit, 240: Overall control unit, 301-302: Camera outside the car, 400: Car, 410: Camera inside the car, 420: Car operation terminal, 430: Door on the car side, 500: Communication base, 601-602 : Car calling terminal, 611 to 612: Hall side door, 1000: Elevator system, 1400: Arrival time estimation device, 401: I / F, 1402: an image memory, 1403: CPU, 1404: RAM, 1405: ROM, 1406: storage unit.

Claims (15)

A passenger number estimating unit that estimates the number of passengers in the car based on images in the car taken by a car camera installed in the elevator car;
A waiting number estimating unit that estimates the number of people waiting for the car based on a video of the getting-on / off location imaged by a camera outside the car installed at the getting-on / off location for getting on / off the elevator out of the floor where the elevator stops;
An elevator information acquisition unit for acquiring operation control information for the elevator;
Based on the number of passengers estimated by the number of passengers estimation unit, the number of standby people estimated by the waiting number estimation unit, and the operation control information acquired by the elevator information acquisition unit, until the car arrives at each floor An arrival time estimation unit for estimating the time of
An elevator arrival time estimation device comprising: The elevator arrival time estimation device is:
The estimated number of passengers estimated by the passenger number estimating unit, the estimated result of the waiting number estimated by the waiting number estimating unit, the operation control information, and the arrival time when the car arrived at each floor are described. It has a past case recording unit that stores past history data,
The arrival time estimation unit
The number of passengers estimated by the number-of-passengers estimation unit and the number of standbys estimated by the waiting number estimation unit, and the number of passengers and the number of standbys in the past history data stored in the past case recording unit When the difference between them is within a predetermined number of people, the arrival time described in the past history data is used as an estimation result for the time until the car arrives at each floor. The elevator arrival time estimation device according to Item 1. The arrival time estimation unit
By limiting only the past history data related to the same time zone on the same day of the week as the comparison target for calculating the difference, only the past history data close to the current status of the elevator and each floor is limited. The elevator arrival time estimation apparatus according to claim 2, wherein the estimation is performed. The passenger number estimating unit
Based on the images in the car taken by the in-car camera, the human attributes in the car are estimated,
The waiting number estimating unit is
Based on the image of the boarding location taken by the camera outside the car, estimating human attributes existing at the boarding location,
The past history data is
Describes the attribute estimation result estimated by the number of passengers estimation unit and the attribute estimation result estimated by the waiting number estimation unit;
The arrival time estimation unit
The difference is calculated for each attribute described by the past history data, and the past history data close to the current situation of the elevator and each floor is specified based on the calculation result. The elevator arrival time estimation device according to claim 2, wherein the arrival time described is the estimation result. The passenger number estimating unit
Based on the image in the car taken by the in-car camera, the foreground of the person in the car and the image in the car is detected, and the number of persons detected is calculated.
For the part hidden by the detected human, the number of humans present in that part is estimated based on the size of the detected foreground,
The total number of persons estimated based on the detected number of people and the size of the detected foreground is the number of passengers,
The waiting number estimating unit is
Based on the image of the boarding location taken by the camera outside the car, detects the foreground of the person existing at the boarding location and the video of the boarding location, and calculates the number of people for the detected human,
For the part hidden by the detected human, the number of humans present in that part is estimated based on the size of the detected foreground,
The elevator arrival time estimation device according to claim 1, wherein a total number of persons estimated based on the detected number of persons and the detected size of the foreground is set as the waiting number. The waiting number estimating unit is
Based on the image of the getting-on / off location photographed by the camera outside the car, a person existing at the getting-on / off location is detected,
When there are a plurality of the elevators in the image of the getting-on / off location, each elevator in the image of the getting-on / off location is based on the positional relationship between the detected person and each elevator in the image of the getting-on / off location. The elevator arrival time estimation apparatus according to claim 1, wherein the waiting number is estimated. The passenger number estimating unit
The image in the car photographed by the in-car camera is divided for each predetermined size, the feature amount is calculated for each section, and the feature amount of the section is shared between the adjacent sections, and then the feature amount To detect a person in the basket using
The waiting number estimating unit is
The image of the boarding / exiting location imaged by the camera outside the car is sectioned for each predetermined size, the feature amount is calculated for each section, the feature amount of the section is shared between the adjacent sections, and the feature amount The elevator arrival time estimation device according to claim 1, wherein a human being present at the boarding / alighting location is detected using. The elevator arrival time estimation device according to claim 1,
One or more elevator cars, a car camera installed in the car,
A camera outside the car installed at a boarding place for getting on and off the elevator on the floor where the elevator stops;
An elevator system comprising: The display unit for displaying an estimation result by the arrival time estimation unit for the car currently installed on each floor and being called at each floor. Elevator system. The display unit
10. The elevator system according to claim 9, wherein an estimation result of the waiting number estimating unit in the intermediate floor until the currently called car arrives in each floor is displayed. The elevator system includes:
An elevator control unit for controlling the operation of the elevator;
The elevator controller is
Calculate the waiting time until the car arrives at each floor based on the arrival time of the car for each floor estimated by the arrival time estimation unit,
The elevator system according to claim 8, wherein the floor on which each car is to be stopped is determined so as to minimize the maximum waiting time among the calculated waiting times on each floor. The elevator controller is
Even if a car is called on the floor on the way the car moves to the floor designated as the destination in the car, the car is not designated as the destination in the car, and If the waiting number on the intermediate floor estimated by the waiting number estimating unit is within the number of persons that can be reliably accommodated by the other car, do not stop the car on the intermediate floor,
When a car is called on the floor on the way the car moves to the floor designated as the destination in the car, and the floor is not designated as the destination in any of the cars, The elevator system according to claim 11, wherein the car to be stopped on the intermediate floor is determined so as to minimize the maximum waiting time. The elevator controller is
When the waiting number estimated by the waiting number estimating unit is a predetermined number or more and the loaded state of the car is a state where the moving speed can be increased, the moving speed of the car is increased. The elevator system according to claim 8. The elevator controller is
The door is closed when the car is stopped on the floor and the door is open, and the waiting number on the floor estimated by the waiting number estimating unit is zero. The elevator system according to claim 8. The elevator controller is
When the car is stopped on the floor and the door is open, and the waiting number on the floor estimated by the waiting number estimating unit is not zero, the passenger number estimating unit is a person in the car. The elevator system according to claim 14, wherein an inquiry is made to the passenger number estimation unit for movement in the car to be detected at the same time, and the door is closed when there is no movement in the car.

Images