Elevator system

Abstract

An elevator system in which users can make a group call without contacting a device. The elevator system is provided with a car assignment unit that assigns a car of an elevator to a landing in accordance with a registered call, and an image analysis unit that detects, from a landing image obtained by photographing a landing, a case where a first user makes a gesture action indicating a target floor, determines the target floor indicated by the gesture action, and detects the number of people in a group of users belonging to the group in a case where a plurality of users including the first user are determined to be one group from the landing image. The car assignment unit generates and registers a call in which the target floor is the destination floor and the number of people in the group board the elevator in a case where the number of people in the group is detected by the image analysis unit and the target floor of the first user is determined.

Description

Technical Field

[0001] This invention relates to elevator systems. Background Technology

[0002] Patent Document 1 discloses an elevator system. The control device of this elevator system receives inputs from the number of users in a group and the destination floor. In this elevator system, the elevator car can be dispatched in a manner that allows all members of the group to ride together.

[0003] Existing technical documents

[0004] Patent documents

[0005] Patent Document 1: International Publication No. 2016 / 142968 Summary of the Invention

[0006] However, in the elevator system described in Patent Document 1, the user operates the system by touching a destination floor registration device located at each floor when entering the number of people in the group and the destination floor. From a hygiene point of view, it is not desirable for the user to come into contact with the device.

[0007] This invention was made to solve the aforementioned problems. The object of this invention is to provide an elevator system that allows users to call the elevator without contacting the device.

[0008] The elevator system of the present invention includes: a car dispatching unit that dispatches an elevator car to a landing based on a registered call; and an image analysis unit that, based on a landing image obtained by capturing images of the landing, detects whether a first user in the landing image performs a gesture indicating a target floor, and determines the target floor indicated by the gesture. If the image analysis unit determines that multiple users, including the first user, belong to a group based on the landing image, it detects the number of users in the group. If the image analysis unit detects the number of users in the group and determines the target floor of the first user, the car dispatching unit generates and registers a call indicating that the target floor is the destination floor and that the group of users takes the elevator.

[0009] Invention Effects

[0010] According to the present invention, in an elevator system, a call button reflecting the number of people in a group and the group's target floor is generated based on an image. Therefore, users can make group calls without contacting the device. Attached Figure Description

[0011] Figure 1 This is a diagram showing the floors of a building equipped with the elevator system of Embodiment 1.

[0012] Figure 2This is a diagram showing a summary of the postures and actions performed by the user of the elevator system according to Embodiment 1.

[0013] Figure 3 This is a diagram showing the content displayed on the display of the elevator system in Embodiment 1.

[0014] Figure 4 This is a diagram showing the interior of the elevator car in Embodiment 1.

[0015] Figure 5 This is a block diagram of the elevator system according to Implementation Method 1.

[0016] Figure 6 This is a flowchart illustrating the operation summary of the control device of the elevator system in Embodiment 1.

[0017] Figure 7 This is a hardware structure diagram of the control device of the elevator system in Implementation Method 1.

[0018] Label Explanation

[0019] 1: Elevator system; 2, 2A, 2B: Car; 3: Control device; 4: Landing; 5A, 5B: Landing door; 6: Landing camera; 7: Display; 8: Car camera; 30: Image judge; 31: Car dispatch controller; 32: Model storage unit; 33: Display control unit; 34: Image analysis unit; 35: Learning and updating unit; 36: Car dispatch unit; 100a: Processor; 100b: Memory; 200: Hardware. Detailed Implementation

[0020] The embodiments for carrying out the invention will be described with reference to the accompanying drawings. Furthermore, in the drawings, identical or equivalent parts are labeled with the same reference numerals. Repetitive descriptions of these parts have been appropriately simplified or omitted.

[0021] Implementation Method 1

[0022] Figure 1 This is a diagram showing the floors of a building equipped with the elevator system of Embodiment 1.

[0023] Figure 1 Elevator system 1 is located in a building not shown. Elevator system 1 has multiple cars 2. For example, elevator system 1 has two cars 2, namely car 2A and car 2B. Control device 3 is located in a machine room not shown. Control device 3 is capable of overall control of elevator system 1.

[0024] Multiple stations 4 are set up on each floor of the building. Figure 1One of several landings 4 is shown. Landing 4 is equipped with landing doors 5A and 5B. Landing door 5A is located at the entrance / exit corresponding to car 2A. Landing door 5B is located at the entrance / exit corresponding to car 2B.

[0025] Multiple landing camera devices 6 are similarly installed at multiple landing stations 4. For example, the landing camera device 6 is installed between landing door 5A and landing door 5B. The landing camera device 6 is capable of capturing landing images of the corresponding landing station 4. The landing camera device 6 is capable of communicating with the control device 3. Multiple displays 7 are similarly installed at multiple landing stations 4 as displays. For example, the displays 7 are liquid crystal displays. The displays 7 are capable of communicating with the control device 3. The displays 7 are capable of displaying information received from the control device 3 to users present at the corresponding landing station 4.

[0026] The landing camera device 6 captures images of users at landing 4. Users perform gestures indicating their target floor. The control device 3 determines the user's target floor based on the landing images captured by the landing camera device 6. The control device 3 selects one of the two elevator cars 2 that is registered to call for that target floor. The control device 3 dispatches the registered elevator car 2 to that landing 4.

[0027] Next, use Figure 2 The user's postures and movements are described.

[0028] Figure 2 This is a diagram illustrating a summary of the postures and movements performed by a user of the elevator system according to Embodiment 1. Additionally, in Figure 2 The diagram of control device 3 is omitted in the text.

[0029] like Figure 2 As shown, at station 4, there are multiple users divided into several groups. Users X1, X2, and X3 belong to group X. For example, group X is a family of three. User Y1 belongs to group Y. Group Y is a one-person group consisting only of user Y1. User Z1 belongs to group Z. Group Z is a one-person group consisting only of user Z1.

[0030] Users use hand gestures to indicate the target floor. In group X, the first user X1, acting as the representative, uses a hand gesture to indicate that the 3rd floor is the target floor. Specifically, the first user X1 uses a gesture with three fingers raised. In group Y, the user Y1, acting as the representative, similarly uses a hand gesture to indicate that the 3rd floor is the target floor. In group Z, the user Z1, acting as the representative, similarly uses a hand gesture to indicate that the 4th floor is the target floor.

[0031] Furthermore, the gesture doesn't have to be a simple finger-raising gesture, but rather a unique gesture specific to each building and each floor. For example, information could be displayed at floor 4 indicating that a call for the elevator to the destination floor can be registered via gesture, along with the floor represented by that gesture. For example, it could also be... Figure 2 The information is displayed on monitor 7, which is not shown in the figure.

[0032] The floor camera device 6 captures images of groups X, Y, and Z present at floor 4. The floor camera device 6 then sends the information of the floor images of groups X, Y, and Z to the control device 3.

[0033] The control device 3 detects the gestures made by the first user X1, user Y1, and user Z1 based on the floor image. Specifically, the control device 3 detects that the target floor indicated by the gesture of the first user X1 is the 3rd floor. The control device 3 detects that the target floor indicated by the gesture of user Y1 is the 3rd floor. The control device 3 detects that the target floor indicated by the gesture of user Z1 is the 4th floor.

[0034] The control device 3 detects users belonging to groups and the number of people in a group based on the floor station image. Specifically, the control device 3 detects that the first user X1 belongs to group X, and that group X is composed of users X1, X2, and X3, based on the floor station image. Similarly, the control device 3 detects that users Y1 and Z1 belong to groups Y and Z, respectively.

[0035] Control device 3 generates an elevator call with the destination floor being the 3rd floor (target floor for user X1) and allowing 3 people (group X) to use the elevator. Control device 3 generates an elevator call with the destination floor being the 3rd floor (target floor for user Y1) and allowing 1 person (group Y) to use the elevator. Control device 3 generates an elevator call with the destination floor being the 4th floor (target floor for user Z1) and allowing 1 person (group Z) to use the elevator. Control device 3 registers the generated elevator calls.

[0036] Next, use Figure 3 Explain the information displayed on monitor 7.

[0037] Figure 3 This is a diagram showing the content displayed on the display screen of the elevator system in Embodiment 1. Figure 3 The illustration of car 2 is omitted in the text.

[0038] like Figure 3 As shown, the display 7 shows the following information, which establishes the corresponding information for the group that generated the elevator call, the target floor of the group, and the car 2 that was registered to call the elevator.

[0039] The leftmost column displays the number of people in each group. Specifically, from top to bottom, it displays the numbers 3, 1, and 1, corresponding to the number of people in groups X, Y, and Z, respectively.

[0040] The second column from the left displays the target floors for each group. Specifically, from top to bottom, the numbers 3, 3, and 4 are displayed as the target floors corresponding to groups X, Y, and Z, respectively.

[0041] The third column from the left shows car 2 that has been registered for a call. Specifically, from top to bottom, the symbols A, A, and B indicate that car 2 has been registered for a call corresponding to groups X, Y, and Z.

[0042] Although not illustrated, for example, display 7 shows the following meaning: Stop the gesture if the target floor is correct, and continue the gesture if the target floor is incorrect.

[0043] If user X1 determines that the target floor shown on display 7 is correct, they stop gesturing. If user X1 determines that the target floor shown on display 7 is incorrect, they continue gesturing to indicate the target floor. Control device 3 detects that user X1 is continuing to gesture. In this case, control device 3 recognizes the gesture that user X1 is continuing to perform, generates a new elevator call, which corresponds to the user's group X and uses the target floor indicated by the gesture as the new target floor.

[0044] Next, use Figure 4 The images captured inside car 2 are explained.

[0045] Figure 4 This is a diagram showing the interior of the elevator car in Embodiment 1. Additionally, in Figure 4 The diagrams of control device 3 and floor station 4 are omitted.

[0046] like Figure 4 As shown, an in-car camera device 8 is installed in car 2A. Additionally, although not shown, a similar in-car camera device is also installed inside car 2B. The in-car camera device 8 is capable of capturing images of the interior of car 2A. The in-car camera device 8 is capable of communicating with the control device 3.

[0047] In car 2A, there are users X1, X2, X3, and W1 who took the elevator from floor 4. The in-car camera device 8 sends the images of users X1, X2, X3, and W1 inside the car to the control device 3.

[0048] Although not illustrated, the control device 3 is in a state where, based on the floor images from the floor camera device 6, it determines that users X1, X2, X3, and W1, who entered car 2A, belong to group X. The control device 3 detects that users X1, X2, X3, and W1 belong to group X based on the images from the car camera device 8 inside the car.

[0049] If user W1 exits the elevator at a floor different from the target floor corresponding to group X, the control device 3 detects this based on the image inside the elevator car. In this case, the control device 3's determination that user W1 belongs to group X is incorrect.

[0050] Furthermore, if user W1 wishes to exit at a floor different from the target floor of group X, it is possible that this different floor is not registered as the destination floor of car 2A. However, in this case, user W1 can still register the desired floor as the destination floor by operating the control panel located inside car 2A. Additionally, if user W1 determines upon boarding car 2A that the desired floor is not registered as the destination floor, they can still exit car 2A. Thus, even if the control device 3 makes an incorrect group determination, car 2A can still operate as usual.

[0051] Next, use Figure 5 The control device 3 will be described.

[0052] Figure 5 This is a block diagram of the elevator system according to Implementation Method 1.

[0053] like Figure 5 As shown, the control device 3 includes an image determination unit 30 and a car dispatch controller 31.

[0054] The image determination unit 30 receives information from images captured by the floor camera 6 and the car camera 8. The image determination unit 30 includes a model storage unit 32, a display control unit 33, an image analysis unit 34, and a learning and updating unit 35.

[0055] The model storage unit 32 stores the learned model. For example, the learned model is pre-stored in the model storage unit 32. The learned model is the model used when analyzing images. Specifically, the learned model outputs a model of the target floor corresponding to the posture and action of the user in the image when an image is input. In addition, the learned model outputs a determination result of whether multiple users in the image belong to the same group when an image is input. At this time, the determination result includes information about the number of people in the group. For example, the learned model is a model that can determine whether multiple users belong to the same group based on the distance between multiple users in the image, the orientation of the user's face, the movement of the user's mouth, etc.

[0056] The display control unit 33 controls the information displayed on the multiple displays 7. Specifically, the display control unit 33 displays the information obtained by the image analysis unit 34 through detection or judgment on the corresponding layer station 4 of the multiple displays 7.

[0057] The image analysis unit 34 obtains the learned model from the model storage unit 32 in advance. The image analysis unit 34 uses the learned model to analyze the received images.

[0058] Based on the floor images from the floor camera device 6, the image analysis unit 34 detects that the first user in the floor images has performed a gesture. The image analysis unit 34 then determines the target floor represented by the gesture.

[0059] The image analysis unit 34 detects cases where multiple users in the floor station image belong to one group. The image analysis unit 34 also detects cases where only one user in the floor station image belongs to one group. If a group is detected, the image analysis unit 34 counts the number of users in that group. If the image analysis unit 34 detects that the first user belonging to that group is performing a gesture, it establishes a correspondence between that gesture and the target floor.

[0060] Based on the floor image, the image analysis unit 34 detects if the first user, whose posture and movement have already been determined, is continuing to perform a posture or movement. If, after a predetermined time has elapsed since the display control unit 33 displayed the first user's target floor information on the display 7, the image analysis unit 34 determines that the first user is continuing to perform a posture or movement. In this case, the image analysis unit 34 determines the target floor indicated by the posture or movement performed by the first user as the new target floor.

[0061] The image analysis unit 34 establishes a correspondence between the group to which the user of car 2 belongs and the group detected from the images from the floor cameras 6, based on the images from the in-car camera device 8. That is, the image analysis unit 34 establishes a correspondence between the group of users in car 2 and the target floor of that group. The image analysis unit 34 also detects, based on the images from the in-car camera device 8, cases where any user belonging to the same group has disembarked at a floor different from the target floor corresponding to that group.

[0062] The learning update unit 35 generates and stores relearning information based on the judgment result of the image analysis unit 34. The relearning information is information that establishes a correspondence between images captured by the floor camera 6 or the car camera 8 and tags indicating that the image is incorrect. The learning update unit 35 performs relearning on the learned model stored in the model storage unit 32 based on the stored relearning information. The learning update unit 35 updates the learned model stored in the model storage unit 32 to the relearned model. Furthermore, the learning update unit 35 can perform relearning and update the learned model at any time.

[0063] As the first example of relearning, when the image analysis unit 34 determines a new target floor, the learning update unit 35 determines that the target floor determined before the new target floor was incorrect. In this case, the learning update unit 35 obtains the floor image at the time when the image analysis unit 34 determined the incorrect target floor as the first misjudged image. The learning update unit 35 generates and stores the following first relearning information as relearning information, which establishes corresponding information for the first misjudged image, the incorrect target floor, and the first tag indicating the incorrect target floor.

[0064] As a second example of relearning, if the image analysis unit 34 detects that a user who was incorrectly identified has descended the stairs on a floor different from the target floor corresponding to their group, the learning update unit 35 determines that the group identification of the user who was incorrectly identified is wrong. In this case, the learning update unit 35 obtains the floor image where the image analysis unit 34 detects that multiple users, including the user who was incorrectly identified, belong to one group as a second incorrect identification image. The learning update unit 35 generates and stores second relearning information as relearning information, which is information that establishes a correspondence between the second incorrect identification image and a second tag indicating that the group identification is incorrect. For example, the learning update unit 35 performs relearning by changing the weights set for each reference used as a basis when determining whether multiple users belong to the same group, such as the distance between users, the orientation of the user's face, and the movement of the user's mouth, based on the second relearning information.

[0065] The car dispatch controller 31 includes a car dispatch unit 36. The car dispatch unit 36 ​​dispatches multiple cars 2. As a control for dispatching cars 2, the car dispatch unit 36 ​​generates a call for a car 2 and registers the generated call for any car 2 among the multiple candidate cars 2.

[0066] Specifically, when the image analysis unit 34 determines the target floor represented by the posture and movement, the car dispatch unit 36 ​​generates a call for the elevator with the target floor as the destination floor and registers any car 2.

[0067] When the image analysis unit 34 detects the number of people in a group and the target floor, the car dispatch unit 36 ​​selects a car 2 from multiple cars 2 that can accommodate that group. At this time, the car dispatch unit 36 ​​compares the number of people in a group shown in the call logs already registered for car 2 with the set passenger capacity for car 2, selecting a car 2 from the multiple candidate cars 2. The car dispatch unit 36 ​​generates a call log with the target floor corresponding to the group detected by the image analysis unit 34 as the destination floor and requests the corresponding group to use the elevator, and registers the selected car 2.

[0068] If the image analysis unit 34 determines a new target floor, the car dispatch unit 36 ​​deletes the elevator call registration corresponding to the group whose new target floor was determined. The car dispatch unit 36 ​​selects a car 2 from multiple cars 2 that can accommodate the group whose new target floor was determined. The car dispatch unit 36 ​​generates an elevator call with the new target floor as the destination floor and allows the corresponding group to use the elevator, and registers the selected car 2.

[0069] After the car dispatch unit 36 ​​registers a call corresponding to a group, if a second user (not the first user) among the multiple users in that group performs a gesture, the image analysis unit 34 detects this gesture based on the landing images from the landing camera device 6. In this case, the image analysis unit 34 detects the additional target floor indicated by the gesture performed by the second user. The image analysis unit 34 removes the second user from the group, reducing the group size by one person. The car dispatch unit 36 ​​generates a call with the additional target floor corresponding to the second user as the destination floor and adds it to any one of the multiple cars 2.

[0070] Next, use Figure 6 The actions performed by the control device 3 based on the floor images from the floor camera device 6 are explained.

[0071] Figure 6 This is a flowchart illustrating the operation summary of the control device of the elevator system in Embodiment 1.

[0072] Figure 6 The flowchart shown illustrates the detection actions for a specific user. When multiple users exist at level 4, the control device 3 performs the detection actions shown in the flowchart for each user.

[0073] In step S01, the control device 3 determines whether the first user performing a gesture has been detected from the floor image of the floor camera device 6.

[0074] If it is determined in step S01 that no first user performing a gesture is detected, the action in step S01 is repeated.

[0075] If, in step S01, a first user performing a gesture is detected, step S02 is performed. In step S02, the control device 3 detects the target floor indicated by the gesture. The control device 3 determines the group to which the first user belongs. The control device 3 generates and registers an elevator call with the detected target floor as the destination floor and requests that the group members use the elevator.

[0076] Then, step S03 is performed. In step S03, the control device 3 causes the display 7 located on the corresponding floor 4 to display the detected information.

[0077] Then, the action of step S04 is performed. In step S04, the control device 3 is in standby mode until a predetermined time has elapsed since the action of step S03.

[0078] Then, the action in step S05 is performed. In step S05, the control device 3 determines whether the first user who detected the gesture is continuing to perform the gesture.

[0079] If it is determined in step S05 that the first user is continuing to perform a gesture, then step S06 is performed. In step S06, the control device 3 detects the new target floor represented by the continued gesture. The control device 3 generates and registers the call corresponding to the new target floor.

[0080] Then, the action in step S07 is performed. In step S07, the control device 3 generates and stores the first relearning information. At this time, the control device 3 can also use the first relearning information to perform relearning.

[0081] If it is determined in step S05 that the first user has not continued the gesture, or if the action in step S07 has been performed, the action in step S08 is performed. In step S08, the control device 3 determines whether the second user belonging to the group has performed a gesture.

[0082] If it is determined in step S08 that the second user has performed a gesture, then step S09 is performed. In step S09, the control device 3 generates and adds a call for the additional target floor corresponding to the second user.

[0083] If the second user is not determined to have performed a gesture in step S08, or if the action in step S09 has been performed, the action in step S10 is performed. In step S10, the control device 3 determines whether the group has boarded the elevator car 2 based on the image from the floor camera device 6.

[0084] If it is determined in step S10 that the group did not ride in car 2, the actions after step S08 are repeated.

[0085] When it is determined in step S10 that the group has boarded car 2, the control device 3 ends the operation of the flowchart.

[0086] According to Embodiment 1 described above, the elevator system 1 includes an image analysis unit 34 and a car dispatch unit 36. The car dispatch unit 36 ​​generates and registers a call to the destination floor corresponding to the group detected by the image analysis unit 34, and to allow the members of the corresponding group to board the elevator. Therefore, the elevator system 1 can generate a call that allows all members of a group to board the elevator without the users of the group coming into contact with equipment such as the landing control panel. As a result, the elevator system 1 can provide users with a hygienic and comfortable environment. Users can use the elevator system 1 without feeling any hygiene concerns.

[0087] Additionally, for example, there exists an elevator system that generates a call for an elevator based on the number of people in a group by detecting the number of people in that group. In this case, the standby space needs to be clearly defined at each floor. However, due to facility limitations, it is sometimes difficult to set up such a standby space. According to the elevator system 1 of this embodiment, it is possible to generate a call for an elevator that allows all members of the group to use the elevator without setting up such a standby space.

[0088] Furthermore, the car dispatching unit 36 ​​selects a car 2 from multiple cars 2 that can accommodate a group of passengers. Therefore, the elevator system 1 can dispatch the most suitable car 2.

[0089] Furthermore, the elevator system 1 also includes a display control unit 33. After the display control unit 33 displays the determination result on the display 7, the elevator system 1 detects that the user is continuing to perform a gesture. In this case, the elevator system 1 determines a new target floor and registers a call based on the new target floor. Therefore, even if the initial target floor determination is incorrect, the elevator system 1 can correct it to a new target floor.

[0090] Furthermore, the elevator system 1 also includes a model storage unit 32, which stores the learned model. The image analysis unit 34 analyzes the images based on the learned model. Therefore, the elevator system 1 can improve the accuracy of target floor determination and group detection.

[0091] Furthermore, elevator system 1 also includes a learning and updating unit 35. Therefore, elevator system 1 can perform relearning on the learned model. As a result, elevator system 1 can improve the accuracy of the learned model.

[0092] Furthermore, the learning update unit 35 generates first relearning information, which includes the first misjudged image. The learning update unit 35 uses the first relearning information to perform relearning on the learned model. Therefore, the elevator system 1 can improve the accuracy of determining the target floor.

[0093] Furthermore, the learning update unit 35 generates second relearning information, which includes a second misjudged image. The learning update unit 35 uses the second relearning information to perform relearning on the learned model. Therefore, the elevator system 1 can improve the detection accuracy of the group.

[0094] Next, use Figure 7 An example of the hardware constituting control device 3 will be explained.

[0095] Figure 7 This is a hardware structure diagram of the control device of the elevator system in Implementation Method 1.

[0096] The functions of the control device 3 can be implemented by a processing circuit. For example, the processing circuit includes at least one processor 100a and at least one memory 100b. For example, the processing circuit includes at least one dedicated hardware 200.

[0097] When the processing circuit includes at least one processor 100a and at least one memory 100b, the functions of the control device 3 are implemented by software, firmware, or a combination of software and firmware. At least one of the software and firmware is described as a program. At least one of the software and firmware is stored in at least one memory 100b. At least one processor 100a reads and executes the program stored in at least one memory 100b, thereby implementing the functions of the control device 3. At least one processor 100a is also referred to as a central processing unit, processing device, arithmetic unit, microprocessor, microcomputer, or DSP. For example, at least one memory 100b is a non-volatile or volatile semiconductor memory such as RAM (Random Access Memory), ROM (Read Only Memory), flash memory, EPROM (Erasable Programmable Read Only Memory), EEPROM (Electrically Erasable Programmable Read Only Memory), disk, floppy disk, optical disk, CD (compact disk), mini disc, DVD (Digital Versatile Disk), etc.

[0098] When the processing circuit has at least one dedicated hardware 200, the processing circuit is implemented, for example, by a single circuit, a composite circuit, a programming processor, a parallel programming processor, an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or a combination thereof. For example, each function of the control device 3 is implemented by a separate processing circuit. For example, each function of the control device 3 is implemented centrally by a processing circuit.

[0099] Regarding the various functions of the control device 3, some can be implemented by dedicated hardware 200, while others can be implemented by software or firmware. For example, the functions implemented by the car dispatcher can be implemented by a processing circuit that is the dedicated hardware 200, and functions other than those implemented by the car dispatcher can be implemented by at least one processor 100a reading and executing a program stored in at least one memory 100b.

[0100] Thus, the processing circuit implements the various functions of the control device 3 through hardware 200, software, firmware, or a combination thereof. Additionally, although not shown, the image determiner 30 can also be installed separately from the control device 3.

Claims

1. An elevator system, wherein, The elevator system has the following features: The car dispatching department dispatches cars to floors based on registered elevator calls; and The image analysis unit, based on the floor image obtained by capturing the floor, detects whether the first user in the floor image has performed a gesture indicating a target floor, and determines the target floor indicated by the gesture. If, based on the layer station images, the image analysis unit determines that multiple users, including the first user, belong to a group, then the number of users belonging to that group within that group is detected. When the image analysis unit detects the number of people in the group and determines the target floor of the first user, the elevator dispatching unit generates and registers an elevator call with the target floor as the destination floor and instructs the group to take the elevator. The elevator system also includes a display control unit, which enables displays located at the landings to show the target floor and the number of people in the group determined by the image analysis unit. If, after a predetermined time has elapsed since the display control unit caused the display to show the target floor and the number of people in the group, the image analysis unit detects that the first user is performing the gesture, it determines the new target floor represented by the gesture. When the image analysis unit determines the new target floor, the car dispatching unit generates and registers an elevator call with the new target floor as the destination floor and the group of people taking the elevator.

2. The elevator system according to claim 1, wherein, The elevator system also includes a model storage unit that stores learned models. These learned models are used to detect, based on image information, whether a person in the image has performed the aforementioned posture or action, and to detect, based on the image, whether multiple people form a group. The image analysis unit detects, based on the learned model stored in the model storage unit, whether the first user in the floor image has performed a gesture representing the target floor, and determines, based on the floor image, that multiple users, including the first user, are a group.

3. An elevator system, wherein, The elevator system has the following features: The car dispatching department dispatches cars to floors based on registered elevator calls; and The image analysis unit, based on the floor image obtained by capturing the floor, detects whether the first user in the floor image has performed a gesture indicating a target floor, and determines the target floor indicated by the gesture. If, based on the layer station images, the image analysis unit determines that multiple users, including the first user, belong to a group, then the number of users belonging to that group within that group is detected. When the image analysis unit detects the number of people in the group and determines the target floor of the first user, the elevator dispatching unit generates and registers an elevator call with the target floor as the destination floor and instructs the group to take the elevator. The elevator system also includes a model storage unit that stores learned models. These learned models are used to detect, based on image information, whether a person in the image has performed the aforementioned posture or action, and to detect, based on the image, whether multiple people form a group. The image analysis unit detects, based on the learned model stored in the model storage unit, whether the first user in the floor image has performed a gesture representing the target floor, and determines, based on the floor image, that multiple users, including the first user, are a group.

4. The elevator system according to claim 2 or 3, wherein, The elevator system also includes a learning and updating unit, which uses relearning information to perform relearning on the learned model and updates the learned model stored in the model storage unit to the relearned learned model. The relearning information is information that establishes a correspondence between the image and the label indicating an error judgment.

5. The elevator system according to claim 4, wherein, If, after a predetermined time has elapsed since the target floor and the number of people in the group were displayed on the monitor located at the landing station, the image analysis unit detects that the first user is performing the gesture, it determines the new target floor represented by the gesture. When the image analysis unit determines the new target floor, the learning update unit generates first relearning information and performs relearning on the learned model based on the first relearning information. The first relearning information establishes corresponding information for the first misjudged image, the target floor, and the label indicating the misjudgment of the target floor. The first misjudged image is a floor image used by the image analysis unit when it determines the target floor.

6. The elevator system according to claim 4, wherein, The image analysis unit, based on the images of the car's interior captured by the camera, detects cases where a user belonging to a group in the car's interior image has mistakenly exited the elevator at a floor different from the target floor corresponding to that group. When the image analysis unit detects that the user who made the incorrect judgment has descended the stairs on a different floor, the learning update unit generates second relearning information and performs relearning on the learned model based on the second relearning information. The second relearning information establishes corresponding information between the second misjudged image and the label indicating the incorrect judgment of the group. The second misjudged image is the floor image that the image analysis unit used when judging the group.

7. The elevator system according to claim 5, wherein, The image analysis unit, based on the images of the car's interior captured by the camera, detects cases where a user belonging to a group in the car's interior image has mistakenly exited the elevator at a floor different from the target floor corresponding to that group. When the image analysis unit detects that the user who made the incorrect judgment has descended the stairs on a different floor, the learning update unit generates second relearning information and performs relearning on the learned model based on the second relearning information. The second relearning information establishes corresponding information between the second misjudged image and the label indicating the incorrect judgment of the group. The second misjudged image is the floor image that the image analysis unit used when judging the group.

8. The elevator system according to any one of claims 1 to 3, 5 to 7, wherein, The car dispatching department registers calls for elevators for the group of people who can ride in one of the multiple candidate cars, including the car in question.

9. The elevator system according to claim 4, wherein, The car dispatching department registers calls for elevators for the group of people who can ride in one of the multiple candidate cars, including the car in question.

Images