Passenger detection system, elevator, and passenger detection method

Abstract

To provide a passenger detection system, an elevator and a passenger detection method which can suppress processing capability of an image analysis part.SOLUTION: A passenger detection system includes a camera installed in a car, and an image analysis part for detecting a passenger in the car from an image imaged by the camera. The image analysis part detects a passenger using a light-weight person detection model (first detection model) at door opening time when a car door of the car is opened, and detects a passenger using a highly accurate person detection model (second detection model) at door closing time when the car door is closed. The highly accurate person detection model can detect a person with hither accuracy than the light-weight person detection model.SELECTED DRAWING: Figure 5

Description

[Technical Field]

[0001] The present invention relates to a passenger detection system, an elevator, and a passenger detection method. [Background technology]

[0002] Generally, there is a known technology that detects people from images or videos captured by a camera and measures the degree of congestion in the elevator. When an elevator that uses this technology detects that the elevator is full, it operates in a way that does not respond to calls from the hall. Also, when it detects that the elevator is empty and no one is getting on or off, the elevator operates by closing the elevator door earlier.

[0003] Patent Document 1 discloses an elevator control system that uses a passenger detection camera. In the elevator control system disclosed in Patent Document 1, a passenger detection camera and a security camera are installed in the elevator car. The passenger detection camera captures the area around the door. The security camera has a wider capture range than the passenger detection camera.

[0004] The elevator control system disclosed in Patent Document 1 creates a composite image by combining footage from a passenger detection camera and footage from a security camera, calculates the passenger occupancy area from the composite image, and determines whether the car is full based on the remaining area obtained by subtracting the passenger occupancy area from the floor area of ​​the car. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Publication No. 2019-85253 Summary of the Invention [Problem to be solved by the invention]

[0006] However, the elevator control system described in Patent Document 1 always uses the same detection model, so even when high accuracy in human detection is not required, it has to perform human detection with accuracy higher than required. Therefore, it is difficult for the elevator control system described in Patent Document 1 to suppress the processing power of the image analysis unit.

[0007] In consideration of the above problems, an object of the present invention is to provide a passenger detection system, elevator, and passenger detection method that can reduce the processing power of the image analysis unit. [Means for solving the problem]

[0008] To solve the above problems and achieve the object of the present invention, a passenger detection system embodying one aspect of the present invention includes a camera installed in a car and an image analysis unit that detects passengers in the car from images captured by the camera. The image analysis unit detects passengers using a first detection model when the car door is opened, and detects passengers using a second detection model when the car door is closed. The second detection model can detect people with higher accuracy than the first detection model. In addition, an elevator that reflects one aspect of the present invention includes a car that moves up and down in a hoistway, a camera installed in the car, and the above-mentioned passenger detection system that has an image analysis unit that detects passengers in the car from images captured by the camera.

[0009] In a passenger detection method embodying one aspect of the present invention, a camera installed in a car captures an image of the interior of the car, and an image analysis unit detects passengers using a first detection model when the car door is open, and detects passengers using a second detection model that can detect people with higher accuracy than the first detection model when the car door is closed. [Effects of the Invention]

[0010] According to the passenger detection system, elevator, and passenger detection method configured as described above, the processing power of the image analysis unit can be reduced. Problems, configurations, and effects other than those described above will become clear from the following description of the embodiments. [Brief explanation of the drawings]

[0011] [Figure 1] 1 is a schematic configuration diagram of an elevator according to a first embodiment. [Figure 2] 1 is a block diagram showing a configuration of a passenger detection system according to a first embodiment. [Figure 3] FIG. 2 is a block diagram showing the functional configuration of a control unit according to the first embodiment. [Figure 4] FIG. 2 is a block diagram showing the functional configuration of an image analysis unit according to the first embodiment. [Figure 5] 5 is a flowchart showing a person detection process according to the first embodiment. [Figure 6] 4 is a flowchart showing a car operation process according to the first embodiment. [Figure 7] FIG. 10 is a block diagram showing the configuration of a passenger detection system according to a second embodiment. [Figure 8] FIG. 10 is a block diagram showing the functional configuration of an image analysis unit according to a second embodiment. [Figure 9] 10 is a flowchart showing a person detection process according to the second embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0012] 1. First embodiment A passenger detection system, an elevator, and a passenger detection method according to a first embodiment will be described below with reference to Figures 1 to 6. Note that common parts in the figures are given the same reference numerals.

[0013] [Elevator] First, the configuration of the elevator according to the first embodiment will be described with reference to FIG. FIG. 1 is a schematic configuration diagram of an elevator according to a first embodiment.

[0014] As shown in Fig. 1, elevator 1 is installed in a hoistway 110 formed within a building structure. Elevator 1 moves up and down within hoistway 110 and includes a car 120 for carrying passengers and cargo, ropes 130, a counterweight 140, a hoisting machine 100, and a control device 190. A machine room 160 is provided at the top of hoistway 110.

[0015] The hoist 100 is disposed in a machine room 160, and raises and lowers the car 120 by winding a rope 130 around it. A deflection sheave 150 is disposed near the hoist 100. The rope 130 is attached to the deflection sheave 150.

[0016] The car 120 is attached to one axial end of the rope 130. The counterweight 140 is attached to the other axial end of the rope 130. Therefore, the car 120 is connected to the counterweight 140 via the rope 130. When the hoisting machine 100 is driven, the car 120 and the counterweight 140 move up and down.

[0017] The elevator car 120 has a car chamber 121 and a car door 122 (see FIG. 2). The car chamber 121 is formed in a hollow, approximately rectangular parallelepiped shape. One side of the car chamber 121 is formed with an entrance / exit for people and luggage to enter and exit. The car door 122 opens and closes the entrance / exit of the car chamber 121.

[0018] The control device 190 corresponds to the control unit according to the present invention. The control device 190 is installed, for example, in the machine room 160. The control device 190 controls the operation of the elevator 1. For example, the control device 190 controls the driving of the hoisting machine 100, the opening and closing of the car door 122 (see FIG. 2) of the car 120, etc. The elevator 1 also has a passenger detection system that detects passengers in the car 120.

[0019] A hall call button (not shown) is installed at the hall of each floor in the building structure. A passenger who wishes to board the elevator 1 presses the hall call button at the hall. The hall call button transmits call information including the floor number for which the hall call button was pressed to the control device 190. The control device 190 registers the hall call in accordance with the call information. The control device 190 then controls the driving of the hoisting machine 100 to move the car 2 to the registered hall floor.

[0020] [Passenger detection system configuration] Next, the passenger detection system of the elevator 1 will be described with reference to FIG. FIG. 2 is a block diagram showing the configuration of the passenger detection system according to the first embodiment.

[0021] 2, the passenger detection system for elevator 1 is composed of an in-car monitoring camera 125 and an image analysis device 200. The in-car monitoring camera 125 corresponds to the camera according to the present invention, and the image analysis device 200 corresponds to the image resolution unit according to the present invention.

[0022] The car interior monitoring camera 125 is installed inside the passenger car 120. The car interior monitoring camera 125 is located on the side wall of the car room 121 opposite the side wall to which the car door 122 is attached. The objective lens of the car interior monitoring camera 125 faces the car door 122. The car interior monitoring camera 125 sends the captured images to the image analysis device 200. Note that the images according to the present invention include video, which is a plurality of images that are taken successively at a constant speed.

[0023] An in-car destination floor registration button 126 is installed inside the car room 121 (car 120). The in-car destination floor registration button 126 is a button that a passenger presses when inputting a destination floor. The in-car destination floor registration button 126 transmits the input destination floor to the control device 190. The control device 190 registers the input destination floor. The control device 190 then controls the drive of the hoisting machine 100 to move the car 2 to the destination floor.

[0024] The image analysis device 200 is disposed on the outside of the upper part of the car room 121. The image analysis device 200 detects passengers in the car 120 from images captured by the car interior monitoring camera 125. The image analysis device 200 transmits the detection results to the control device 190. The control device 190 controls the raising and lowering operation of the car 120 and the opening and closing operation of the car door 122 according to the detection results of the image analysis device 200.

[0025] [Control device functional configuration] Next, the functional configuration of the control device 190 will be described with reference to FIG. FIG. 3 is a block diagram showing the functional configuration of the control device 190.

[0026] As shown in FIG. 3, the control device 190 has a communication unit 191, an elevator operation control unit 192, a call registration unit 194, a destination floor registration unit 195, an elevation control unit 196, a door opening / closing control unit 197, and a door opening / closing detection unit 198.

[0027] The communication unit 191 receives various types of information sent from the image analysis device 200, the car 120, the hoisting machine 100, etc. Examples of the various types of information include the detection results of the image analysis device 200, car door open / close information, destination floor information, call information, and the rotation speed of the hoisting machine 100. The communication unit 191 sends the received various types of information to the elevator operation control unit 192.

[0028] The elevator operation control unit 192 controls the operation of the elevator 1. The elevator operation control unit 192 performs various calculations to control the overall operation of the elevator 1, including the ascent and descent speed of the car 120 and the opening and closing operation of the car doors. The elevator operation control unit 192 sends the calculation results to the call registration unit 194, the destination floor registration unit 195, the elevation control unit 196, and the door opening and closing control unit 197.

[0029] The call registration unit 194 registers a platform call based on the call information sent from the platform call button. Platform call registration is not limited to being based on pressing a platform call button, but may be based on information sent from an external device such as a mobile terminal carried by a passenger. The call registration unit 194 sends the platform floor registered in the platform call registration to the elevator operation control unit 192.

[0030] The destination floor registration unit 195 registers the destination floor based on information sent from the in-car destination floor registration button 126. Note that the registration of the destination floor is not limited to being based on pressing the in-car destination floor registration button 126, but may also be based on information sent from an external device such as a mobile terminal carried by a passenger. The destination floor registration unit 195 sends the registered destination floor to the elevator operation control unit 192.

[0031] The elevator operation control unit 192 generates a drive command for the hoisting machine 100 in accordance with the registered destination floor and landing floor. Then, the elevator operation control unit 192 sends the drive command for the hoisting machine 100 to the lifting / lowering control unit 196. The lifting / lowering control unit 196 controls the drive of the hoisting machine 100 in accordance with the drive command for the hoisting machine 100 sent from the elevator operation control unit 192. As a result, the car 2 moves to the landing floor or destination floor.

[0032] The door open / close detection unit 198 detects whether the door of the car 2 is fully open or fully closed based on the detection result of an open / close sensor (or open / close switch) provided in the car 2. The door open / close detection unit 198 sends the detection result to the elevator operation control unit 192. The elevator operation control unit 192 generates a drive command for the car door 122 in accordance with the detection result of the door open / close detection unit 198.

[0033] The door opening / closing control unit 197 controls the opening and closing of the car door 122 in response to a drive command for the car door 122 sent from the elevator operation control unit 192. The landing doors on each floor in the building structure open and close in conjunction with the car door 122. The door opening / closing control unit 197 opens the car door 122 of the car 120 that has arrived at the landing floor, the destination floor, or both. The door opening / closing control unit 197 closes the car door 122 when a predetermined time has elapsed since the car door 122 was opened.

[0034] [Functional configuration of image analysis device] Next, the functional configuration of the image analyzing device 200 will be described with reference to FIG. FIG. 4 is a block diagram showing the functional configuration of the image analyzing device 200.

[0035] 4, image analyzing device 200 includes image acquiring unit 201, door signal acquiring unit 202, detection result responding unit 203, and human detecting unit 204. Furthermore, image analyzing device 200 includes light human detection model storing unit 205, high-accuracy human detection model storing unit 206, and detection area coordinate storing unit 207.

[0036] The image acquisition unit 201 acquires an image (video) captured by the car interior monitoring camera 125. The image acquisition unit 201 sends the acquired image to the person detection unit 204. The door signal acquisition unit 202 acquires a door signal sent by the control device 190. The door signal is a signal that indicates whether the car door 122 is open or closed. The door signal acquisition unit 202 sends the acquired door signal to the person detection unit 204. The detection result response unit 203 sends the detection result of the person detection unit 204 to the control device 190.

[0037] The lightweight person detection model storage unit 205 stores a lightweight person detection model and parameter values ​​(configuration) used for the lightweight person detection model. The lightweight person detection model corresponds to the first person detection model according to the present invention. The lightweight person detection model is a machine learning model that detects people appearing in a specific area (the entire area in this embodiment) of an image or video.

[0038] The lightweight human detection model detects whether or not a person is present. The lightweight human detection model has a simple structure and requires a small number of parameters in order to relatively shorten the time required to detect whether or not a person is present.

[0039] The high-precision person detection model storage unit 206 stores a high-precision person detection model and parameter values ​​(configuration) used for the high-precision person detection model. The high-precision person detection model corresponds to the second person detection model according to the present invention. The high-precision person detection model is a machine learning model that detects people appearing in a specific area (the entire area in this embodiment) of an image or video.

[0040] The high-precision person detection model detects whether or not a person is present and how many people there are. The high-precision person detection model has a more complex structure than the lightweight person detection model and requires a larger number of parameters. The high-precision person detection model can detect people with higher accuracy than the lightweight person detection model, and can, for example, detect the number of people with higher accuracy.

[0041] The detection area coordinate storage unit 207 stores the coordinates of the area in which a person is to be detected. In this embodiment, people are detected in the entire area of ​​the image captured by the car interior monitoring camera 125. Note that the area in which a person is to be detected according to the present invention may be set to a specific area in the image captured by the car interior monitoring camera 125.

[0042] The person detection unit 204 determines whether the door is open or closed based on the door signal. The person detection unit 204 determines the person detection model to use depending on whether the door is open or closed. The person detection unit 204 sends the detection result to the detection result response unit 203.

[0043] When the door is opened, the human detection unit 204 uses the lightweight human detection model and its parameter values ​​to detect a human within the entire area of ​​the image captured by the car interior monitoring camera 125. This allows the human detection unit 204 to detect the presence or absence of a human relatively quickly when the door is opened. Furthermore, the processing capacity of the image analysis device 200 can be reduced compared to when a high-precision human detection model is used to detect the presence or absence of a human in approximately the same amount of time.

[0044] When the doors are closed, the person detection unit 204 uses the high-precision person detection model and its parameter values ​​to detect people within the entire area of ​​the image captured by the car interior monitoring camera 125. Then, the person detection unit 204 sends the detection result to the detection result response unit 203. This allows the person detection unit 204 to detect the number of people with high precision when the doors are closed. The person detection unit 204 can detect whether the car 120 is full or not when the doors are closed.

[0045] [Person detection processing] Next, the person detection process performed by the image analyzing device 200 will be described with reference to FIG. FIG. 5 is a flowchart showing the person detection process performed by the image analyzing device 200.

[0046] First, the image analyzing device 200 checks whether the car door 122 is open or closed based on the door signal sent from the control device 190 (S1). Next, the image analyzing device 200 determines whether the car door 122 is open or not (S2).

[0047] In step S2, when it is determined that the car door 122 is open (YES in S2), the image analyzing device 200 uses the light-weight person detection model to perform person detection when the door is open (S3).

[0048] Next, the image analyzing device 200 determines whether or not a person is present in the image area (S4). In step S4, when it is determined that a person is present in the image area (YES determination in S4), the image analyzing device 200 ends the person detection process.

[0049] In step S4, when it is determined that there is no person in the image area (if S4 is determined as NO), the image analyzing device 200 notifies the control device 190 that the car 120 is unoccupied (S5). After processing in step S5, the image analyzing device 200 ends the person detection process.

[0050] On the other hand, when it is determined in step S2 that the car door 122 is not open (the door is closed) (if S2 is determined to be NO), the image analysis device 200 uses a high-precision person detection model to detect a person when the door is closed (S6).

[0051] Next, the image analyzing device 200 determines whether the number of people in the image area is equal to or greater than a threshold (S7). In step S7, when it is determined that the number of people in the image area is not equal to or greater than the threshold (NO in S7), the image analyzing device 200 ends the person detection process.

[0052] In step S7, when it is determined that the number of people in the image area is equal to or greater than the threshold value (YES determination in S7), the image analyzing device 200 notifies the control device 190 that the car 120 is full (S8). After processing in step S8, the image analyzing device 200 ends the person detection process.

[0053] In this embodiment, when the car door 122 is opened, it is determined whether or not the car 120 is unoccupied. This makes it possible to determine whether or not the car 120 is unoccupied when passengers get on or off. As a result, it is possible to quickly detect that the car 120 is unoccupied.

[0054] Furthermore, the control device 190 can change the timing of closing the car door 122 depending on whether the car 120 is empty. In order to change the timing of closing the car door 122, it is desirable to detect as quickly as possible whether the car 120 is empty (0 people or 1 or more people) when the door is opened. In this embodiment, a lightweight person detection model is used to detect people when the door is opened, so it is possible to detect relatively quickly whether the car 120 is empty while suppressing the processing capacity of the image analysis device 200.

[0055] In this embodiment, it is determined whether the car 120 is full when the car door 122 is closed. This makes it possible to determine whether the car 120 is full when no passengers are getting on or off, thereby improving the accuracy of determining whether the car is full.

[0056] Furthermore, the control device 190 can change the stopping floor of the car 120 depending on whether the car is full or not. In this embodiment, since a high-precision person detection model is used to detect people when the door is closed, the number of people in the image area can be detected with high accuracy even if people (passengers) in the image area overlap. Furthermore, since a longer detection time can be secured when the door is closed compared to when the door is open, the processing capacity of the image analyzing device 200 can be reduced even when a high-precision person detection model is used.

[0057] [Car operation processing] Next, the car operation processing performed by the control device 190 will be described with reference to FIG. FIG. 6 is a flowchart showing the car operation process performed by the control device 190.

[0058] First, the control device 190 checks the open / closed state of the car door 122 based on the detection result of the door open / close detection unit 198 (S11). Next, the control device 190 determines whether the car door 122 is open or not (S12).

[0059] In step S12, when it is determined that the car door 122 is open (if S12 is determined as YES), the control device 190 determines whether or not the car 120 is unoccupied based on the detection result of the image analysis device 200 (S13).

[0060] When it is determined in step S13 that the car 120 is unoccupied (when S13 is determined as YES), the control device 190 closes the car door 122 (S14). Normally, the control device 190 closes the car door 122 after a timer expires. The timer is set to a predetermined time. However, when the car 120 is unoccupied, the control device 190 closes the car door 122 at an earlier timing than usual without waiting for the timer to expire. After processing step S14, the control device 190 terminates the car operation processing.

[0061] In step S13, when it is determined that the car 120 is not unoccupied (when S13 is determined as NO), the control device 190 closes the car door 122 after the timer expires (S15). A predetermined time is set in the timer. Therefore, when the predetermined time has elapsed since the car door 122 was opened, the control device 190 closes the car door 122. After processing step S15, the control device 190 ends the car operation processing.

[0062] On the other hand, when it is determined in step S12 that the car door 122 is not open (the door is closed) (if S12 is determined as NO), the control device 190 determines whether the car 120 is full or not based on the detection result of the image analysis device 200 (S16).

[0063] In step S16, when it is determined that the car 120 is full (when S16 is determined as YES), the control device 190 determines to have the car 120 pass through the call-registered landing floors on the way to the next destination floor (S17). After processing step S17, the control device 190 ends the car operation processing.

[0064] In step S16, when it is determined that the car 120 is not full (when S16 is determined as NO), the control device 190 stops the car 120 at the boarding floor corresponding to the call registration (S18). That is, the control device 190 determines to stop the car 120 at a call-registered boarding floor that is on the way to the next destination floor. After processing step S18, the control device 190 ends the car operation processing.

[0065] In this embodiment, if the car 120 is unoccupied when the door is opened, the car door 122 is closed earlier than usual. This reduces unnecessary waiting time and allows the car 120 to quickly move to the floor corresponding to the next destination floor registration or call registration. Furthermore, if there is a passenger getting off the car 120, the car door 122 is not closed earlier than usual by mistake.

[0066] 2. Second embodiment Next, a passenger detection system, an elevator, and a passenger detection method according to a second embodiment will be described with reference to Figures 7 to 9. Note that common parts in each figure are given the same reference numerals.

[0067] The passenger detection system according to the second embodiment differs from the passenger detection system according to the first embodiment in the image analysis device of the passenger detection system. Therefore, the configuration of the passenger detection system according to the second embodiment, the image analysis device, and the person detection process will be described here, and a description of the configuration that overlaps with the first embodiment will be omitted.

[0068] [Passenger detection system configuration] First, a passenger detection system according to a second embodiment will be described with reference to FIG. FIG. 7 is a block diagram showing the configuration of a passenger detection system according to the second embodiment.

[0069] 7, the passenger detection system according to the second embodiment is made up of an in-car monitoring camera 125 and an image analysis device 200 B. The image analysis device 200 B corresponds to the image resolution unit according to the present invention.

[0070] An in-car destination floor registration button 126 and a signage device 127 are installed in the car room 121 of the car 120B. The signage device 127 is a device that transmits advertisements and information using images, videos, and sounds. The signage device 127 has, for example, a display using a liquid crystal or an LED (Light Emitting Diode), and a speaker.

[0071] The image analysis device 200B is disposed on the upper outside of the car room 121. The image analysis device 200B detects the presence or absence of passengers in the car 120, the number of passengers, the attributes of the passengers, etc. from the image captured by the car interior monitoring camera 125. The image analysis device 200B transmits the detection results to the control device 190.

[0072] The control device 190 controls the raising and lowering operation of the car 120 and the opening and closing operation of the car door 122 according to the detection results of the image analysis device 200B. The control device 190 manages content such as advertisements and information to be displayed on the signage device 127 of the car 120. The control device 190 controls the signage device 127 to transmit advertisements and information according to passengers.

[0073] Control device 190 sends passenger attributes from the detection results of image analysis device 200B to signage content server 300 provided in the architectural structure. Signage content server 300 manages content such as advertisements and information to be displayed on signage devices (not shown) installed in the architectural structure. Signage content server 300 controls the signage devices (not shown) to transmit advertisements and information.

[0074] [Functional configuration of image analysis device] Next, the functional configuration of the image analyzing device 200B will be described with reference to FIG. FIG. 8 is a block diagram showing the functional configuration of the image analyzing device 200B.

[0075] 8, image analyzing device 200B has image acquiring unit 201, door signal acquiring unit 202, detection result responding unit 203, and human detecting unit 204. Furthermore, image analyzing device 200B has light human detection model storing unit 205, high-accuracy human detection model storing unit 206, detection area coordinate storing unit 207, and attribute detection model storing unit 208.

[0076] The attribute detection model storage unit 208 stores attribute detection models and parameter values ​​(configurations) used for the attribute detection models. The attribute detection models are machine learning models that detect the attributes of people appearing in a specific area (the entire area in this embodiment) of an image or video. Examples of attributes detected by the attribute detection models include age (age group), gender, occupation, emotion, and race.

[0077] When the door is open, the person detection unit 204 uses the light-weight person detection model and its parameter values ​​to detect people within the entire area of ​​the image captured by the car interior monitoring camera 125. When the door is closed, the person detection unit 204 uses the high-precision person detection model and its parameter values ​​to detect people within the entire area of ​​the image captured by the car interior monitoring camera 125. Then, when the door is closed, the person detection unit 204 uses the attribute detection model and its parameter values ​​to detect the attributes of people within the entire area of ​​the image captured by the car interior monitoring camera 125.

[0078] [Person detection processing] Next, the person detection process performed by the image analyzing device 200B will be described with reference to FIG. FIG. 9 is a flowchart showing the person detection process performed by the image analyzing device 200B.

[0079] First, the image analyzing device 200B checks the open / closed state of the car door 122 based on the door signal sent from the control device 190 (S21). Next, the image analyzing device 200B determines whether the car door 122 is open or not (S22).

[0080] In step S22, when it is determined that the car door 122 is open (YES in S22), the image analyzing device 200B uses the light-weight person detection model to perform person detection when the door is open (S23).

[0081] Next, image analyzing device 200B determines whether or not a person is present in the image area (S24). When it is determined in step S24 that a person is present in the image area (YES determination in S24), image analyzing device 200B ends the person detection process.

[0082] In step S24, when it is determined that there is no person in the image area (if S24 is determined as NO), the image analyzing device 200B notifies the control device 190 that the inside of the car 120 is unoccupied (S25). After processing in step S25, the image analyzing device 200B ends the person detection processing.

[0083] On the other hand, when it is determined in step S22 that the car door 122 is not open (the door is closed) (if S22 is determined to be NO), the image analyzing device 200B determines whether or not this is the first time a person has been detected since the door was opened (S26). When the car door 122 is closed, the image analyzing device 200B performs the person detection process two or more times.

[0084] In step S26, when it is determined that this is the first person detection after the door is opened (YES in S26), the image analyzing device 200B uses the high-accuracy person detection model to perform person detection when the door is closed (S27).

[0085] Next, image analyzing device 200B determines whether the number of people in the image area is equal to or greater than a threshold (S28). When it is determined in step S28 that the number of people in the image area is not equal to or greater than the threshold (NO in S28), image analyzing device 200B ends the person detection process.

[0086] In step S28, when it is determined that the number of people in the image area is equal to or greater than the threshold value (YES determination in S28), the image analyzing device 200B notifies the control device 190 that the car 120 is full (S29). After processing in step S29, the image analyzing device 200B ends the person detection process.

[0087] In step S26, when it is determined that this is not the first person detection since the door was opened (if S26 is determined to be NO), the image analyzing device 200B uses the attribute detection model to detect the attributes of the person in the elevator 120 (S30).

[0088] Next, image analyzing device 200B creates statistics of the person attributes detected in step S30 and transmits them to control device 190 (S31). After the process of step S31, image analyzing device 200B ends the person detection process.

[0089] Based on the statistics of person attributes, control device 190 determines the type of content, such as advertisements or information, to be displayed on signage device 127. Then, control device 190 controls signage device 127 to transmit the determined content. Control device 190 also sends the statistics of person attributes to signage content server 300.

[0090] When the car door 122 is closed, the number of passengers in the car 120 does not change. Therefore, there is ample time to perform person detection. Therefore, the image analyzing device 200B detects person attributes after completing person detection using the high-precision person detection model. This allows the elevator 1 and the building structure in which the elevator 1 is installed to provide services (for example, advertisements and information transmission) based on statistics of person attributes.

[0091] The above describes embodiments of the passenger detection system, elevator, and passenger detection method, including their effects. However, the passenger detection system, elevator, and passenger detection method of the present invention are not limited to the above-described embodiments, and various modifications are possible within the scope of the invention as defined in the claims. Furthermore, the above-described embodiments have been described in detail to clearly explain the present invention, and the present invention is not necessarily limited to those having all of the described configurations.

[0092] In the above-described embodiment, the control device 190 and the image analysis devices 200 and 200B are provided separately. However, the control unit and image analysis unit according to the present invention may be configured to be included in a single control device.

[0093] In addition, in the present invention, it is possible to replace part of the configuration of one embodiment with the configuration of another embodiment, or to add the configuration of another embodiment to the configuration of one embodiment.In addition, it is possible to add, delete, or replace part of the configuration of each embodiment with another configuration. [Explanation of symbols]

[0094] 1...Elevator, 100...Hoisting machine, 110...Hoistway, 121...Cab room, 122...Cab door, 125...Internal surveillance camera, 126...Internal destination floor registration button, 127...Signage device, 130...Rope, 140...Counterweight, 150...Bearing wheel, 160...Machine room, 190...Control device (control unit), 200, 200B...Image analysis device (image analysis unit), 201...Image acquisition unit, 202...Door signal acquisition unit, 203...Detection result response unit, 204...Person detection unit, 205...Lightweight person detection model storage unit, 206...High-precision person detection model storage unit, 207...Detection area coordinate storage unit, 208...Attribute detection model storage unit, 300...Signage content server

Claims

1. A camera installed inside the car, An image analysis unit that detects passengers in the elevator car from the image captured by the camera, the image analysis unit detects a passenger using a first detection model when a car door of the car is opened, and detects a passenger using a second detection model when the car door is closed; The second detection model is capable of detecting a person with higher accuracy than the first detection model. Passenger detection system.

2. The parameters used in the first detection model are different from the parameters used in the second detection model. The occupant detection system of claim 1 .

3. The image analysis unit detects passengers using the second detection model when the door is closed, and if there is at least one passenger, detects attributes of the person using an attribute detection model. The occupant detection system of claim 1 .

4. The image analysis unit detects passengers using the second detection model when the door is closed, and determines that the elevator car is full if the number of passengers is equal to or greater than a threshold. The occupant detection system of claim 1 .

5. The image analysis unit detects a passenger using the first detection model when the door is opened, and determines that the elevator car is unoccupied if there is no passenger. The occupant detection system of claim 1 .

6. An elevator comprising: a car that moves up and down in a hoistway; a camera installed in the car; and a passenger detection system that includes an image analysis unit that detects passengers in the car from images captured by the camera, the image analysis unit detects a passenger using a first detection model when a car door of the car is opened, and detects a passenger using a second detection model when the car door is closed; The second detection model is capable of detecting a person with higher accuracy than the first detection model. Elevator.

7. A camera installed in the car captures an image of the inside of the car, An image analysis unit detects a passenger using a first detection model when a car door of the car is opened, and detects a passenger using a second detection model that can detect a person with higher accuracy than the first detection model when the car door is closed. Passenger detection method.

Images