Elevator system and elevator control method

Abstract

To reduce the load on an image analysis of a camera installed in an elevator.SOLUTION: An elevator system comprises an object detection processing part 23 to detect the presence / absence of a person or an object in a car in the analysis of the image captured by a camera 12 installed in a car, a person detection processing part 22 to detect the attribute of persons in a car in the analysis of the image captured by the camera 12, an image analysis system switching part 21 to selectively execute the detection in the object detection processing part 23 and the detection in the person detection processing part 22, and an image analysis result determination part 24 to determine either image analysis result of the detection in the object detection processing part 23 and of the detection in the person detection processing part 22. The notification to an elevator operation control or to the other system connected to the elevator system is given based on the determined result by the image analysis result determination part 24.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to an elevator system and an elevator control method. [Background technology]

[0002] In recent years, cameras have been installed in elevators and at the landing area to detect elevator users from images captured by the cameras. By equipping an elevator with a camera, it is possible to, for example, detect whether the elevator is crowded, detect whether there are no passengers, detect abnormalities among passengers, and check the state of the elevator when an abnormality occurs, from the captured images.

[0003] The elevator can also collect detailed information about elevator usage by analyzing images captured by the camera to detect the attributes of elevator users, such as age and gender. Furthermore, the elevator can use the camera as a security measure to restrict elevator users by performing facial recognition from images captured by the camera.

[0004] Patent Document 1 describes a technology that uses an image analysis algorithm to detect users who intend to board an elevator by installing a camera above the elevator door and analyzing the image to determine which users intend to board the elevator. Patent Document 1 also describes how the image analysis algorithm installed in the image analysis engine that analyzes the image from the camera above the elevator door is switched depending on the detection of the door opening direction. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] JP 2020-152528 A Summary of the Invention [Problem to be solved by the invention]

[0006] However, an image analysis device that analyzes camera images is capable of advanced detection, such as identifying an individual through facial recognition, but in order to perform advanced detection such as facial recognition, highly advanced image analysis processing is required, such as constantly detecting faces in camera images and analyzing the detected faces with high accuracy. If an image analysis device that performs advanced analysis processing is built into an elevator car or control device, the elevator system will require such an advanced device, which creates a problem of complicating the configuration of the elevator system. In addition, if the elevator system constantly performs advanced image analysis processing, the burden of the power consumption required for this will also increase, resulting in increased power consumption of the elevator system.

[0007] Patent document 1 describes a technology that, when an elevator user is detected from a camera image, determines whether the user intends to board the elevator based on the direction in which the doors are opening, and switches between this algorithm and an image analysis algorithm that identifies users who intend to board the elevator. The technology for switching image analysis algorithms described in Patent Document 1 is an algorithm for analyzing images of users (people), and any algorithm requires relatively advanced image analysis. For this reason, the switching process described in Patent Document 1 limits the reduction in the load on the elevator control device, and there is a demand for further reduction in the load on the device and reduction in power consumption.

[0008] In view of the above, an object of the present invention is to provide an elevator system and an elevator control method that can reduce the burden of analyzing camera images. [Means for solving the problem]

[0009] In order to solve the above problems, for example, the configurations described in the claims are adopted. The present application includes multiple means for solving the above-mentioned problems. One example is an elevator system that performs processing according to an image captured by a camera installed inside the elevator car, and includes an object detection processing unit that detects the presence or absence of a person or object inside the elevator car by analyzing the image captured by the camera, a person detection processing unit that detects attributes of the person inside the elevator car by analyzing the image captured by the camera, an image analysis method switching unit that selectively executes detection in the object detection processing unit and detection in the person detection processing unit, and an image analysis result determination unit that determines the image analysis result of either the detection in the object detection processing unit or the detection in the person detection processing unit, and based on the result determined by the image analysis result determination unit, the elevator operation is controlled or a notification is sent to another system connected to the elevator system. Effect of the Invention

[0010] According to the present invention, by selectively performing an object detection processing section and a human detection processing section, appropriate detection is possible while reducing the burden of image analysis. Problems, configurations and effects other than those described above will become apparent from the following description of the embodiments. [Brief description of the drawings]

[0011] [Figure 1] 1 is a diagram showing a schematic configuration of an entire elevator system according to an embodiment of the present invention; [Diagram 2] 1 is a block diagram showing a configuration example of an elevator system according to an embodiment of the present invention. FIG. [Diagram 3] 1 is a block diagram showing an example of a hardware configuration of a control system according to an embodiment of the present invention. [Figure 4] 3 is a flowchart showing a processing example (example 1) of the elevator system according to the embodiment of the present invention. [Diagram 5] 5 is a flowchart showing a processing example (example 2) of the elevator system according to the embodiment of the present invention. [Figure 6]5 is a flowchart showing a processing example (example 3) of the elevator system according to the embodiment of the present invention. [Figure 7] 10 is a flowchart showing a processing example (example 4) of the elevator system according to the embodiment of the present invention. [Figure 8] 10 is a flowchart showing a processing example (example 5) of the elevator system according to the embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] An elevator system according to one embodiment of the present invention (hereinafter, referred to as "this embodiment") will be described below with reference to the accompanying drawings.

[0013] [Overall schematic configuration of elevator system] First, the schematic configuration of the elevator system of this embodiment will be described with reference to FIG. The elevator system of this example includes a car 10 that travels (moves up and down) under the control of an elevator control device 100. The car 10 runs by being driven by a hoist (not shown) based on the elevator control device 100. Note that in Fig. 1, a mechanism for running the car 10, such as the hoist, is omitted. The car 10 is equipped with doors 11 which open and close when the car arrives at a landing on each floor. Further, the car 10 is equipped with a camera 12 and a lighting device 13. The camera 12 is installed on the ceiling or the like of the car 10, and captures the state inside the car 10 as video or still images. The camera 12 is installed on the ceiling or wall at the far end away from the door 11 of the car 10, and is capable of capturing images of not only the inside of the car 10, but also the door 11. Therefore, in the image captured by the camera 12, almost the entire inside of the car 10 is captured from the ceiling side, and the state of the passenger getting on the car 10 through the door 11 (such as the face of the passenger) is also captured.

[0014] The image captured by the camera 12 is analyzed by an image analysis device 20 installed in the car 10. The image analysis device 20 is installed, for example, on the upper part (above the ceiling) of the car 10. However, providing the image analysis device 20 in the car 10 is just one example, and the image analysis device 20 may be installed together with the elevator control device 100, for example, in an elevator machine room or a pit. Alternatively, the image analysis device 20 may be integrated with the camera 12. The lighting device 13 is turned on while the elevator is providing passenger transportation service, illuminating the inside of the elevator car 10. Conversely, while the passenger transportation service is stopped (paused), the lighting by the lighting device 13 is stopped.

[0015] The results of the analysis by the image analysis device 20 are transmitted together with the captured image to the elevator control device 100. The elevator control device 100 determines the usage status of the car 10 based on the image analysis results by the image analysis device 20, and controls the operation of the elevator based on this determination result. The results of the analysis by the image analysis device 20 are also transmitted to the monitor system 200 together with the captured image. A plurality of external systems (a first external system 210, a second external system 220, a third external system 230) are connected to the monitor system 200, and the image analysis results are used in each of the external systems.

[0016] [Elevator control device and image analysis device configuration] Fig. 2 shows the configuration of each part of the elevator system of this example. Note that Fig. 2 shows only the configuration of the elevator control device 100 related to the processing of images captured by the camera 12, and omits other configurations of the elevator control device 100. The elevator control device 100 includes a door open state acquisition unit 111, an operation state acquisition unit 112, a door open state determination unit 113, an operation state determination unit 114, an elevator state management unit 115, and an elevator control processing unit 116.

[0017] The door open state acquisition unit 111 acquires data on the door open state of the door 11 from a switch or a sensor (not shown) installed on the door 11 of the car 10. The operating state acquisition unit 112 acquires data on the operating state (running, stopped, etc.) from the operating state detection unit 101 of the car 10. The door open state determination unit 113 determines the open state of the door 11 based on the data acquired by the door open state acquisition unit 111 . The operating state determination unit 114 determines the operating state of the car 10 based on the data acquired by the operating state acquisition unit 112.

[0018] The elevator status management unit 115 manages elevator operation based on the door open status determined by the door open status determination unit 113 and the operation status determined by the operation status determination unit 114. Furthermore, the elevator status management unit 115 generates an image analysis method switching command based on the door open status and the operation status, and sends the generated image analysis method switching command to the image analysis device 20, or sends the door open status, operation status, etc. to the image analysis device 20 at regular intervals or when the status changes. The elevator control processing unit 116 sends an operation control command to the operation control unit 102 of the car 10. The operation control unit 102 controls the operation of the car 10. The operation control command includes a command to open and close the door 11 as well as a command to run the car 10. The command to open and close the door 11 is also transmitted to the door open state acquisition unit 111. When the elevator control processing unit 116 generates an operation control command for the operation control unit 102, it reflects the image analysis result transmitted from the image analysis device 20. For example, when the image analysis device 20 determines that the door 11 is in an open state at a stop floor according to the number of passengers in the car 10, the time for which the door open state continues is changed. For example, when the number of passengers is large, the door open time at the stop floor due to the stop floor registration operation is made longer than usual. The process of changing the time for which the door open state continues may be further changed according to the age group of the passengers and the discrimination of wheelchair users. For example, when it is determined that an elderly person, a child, or a person carrying luggage is aboard and therefore requires longer than the normal door-opening time to board or disembark, the door-opening time at the stopping floor is made longer than normal.

[0019] The image analysis device 20 installed in the elevator 10 acquires the captured image output by the image management unit 12a of the camera 12, and performs analysis of the acquired image. The image analysis device 20 includes an image analysis method switching unit 21 , a human detection processing unit 22 , an object detection processing unit 23 , an image analysis result determination unit 24 , and an image analysis result transmission unit 25 . The image analysis method switching unit 21 executes image analysis method switching processing based on the door state and operation state, which are signals from the elevator control device 100. The switching processing here is switching processing between human detection processing in the human detection processing unit 22, which will be described next, and object detection processing in the object detection processing unit 23. The human detection processing in the human detection processing unit 22 and the object detection processing in the object detection processing unit 23 are configured to selectively perform only one of them.

[0020] The person detection processing unit 22 executes a process of detecting person attributes from an image captured by the camera 12 through image analysis processing. This process of detecting person attributes detects, for example, the person's general age, gender, and the like. Furthermore, if the person detection processing unit 22 can detect a face from an image captured by the camera 12, it performs a process of identifying the person through face recognition processing. Furthermore, the person detection processing unit 22 also detects wheelchair users, stroller users, users of assistive devices such as canes, and the like through image analysis. This person detection process using image analysis may be performed by learning the image data of the user to generate an attribute analysis model, or by comparing the data with stored data in a database prepared in advance. This person detection process is a process for detecting the details of a person's attributes, and requires a relatively large amount of calculation processing. Also, the image data used to generate the model may be a created CG image instead of actual image data.

[0021] The object detection processing unit 23 performs processing to detect the presence or absence of objects from the image captured by the camera 12 through image analysis processing. The objects in this case include people. Therefore, the object detection processing unit 23 detects objects in the elevator 10 and determines whether they are passengers or not. In other words, it detects whether there are passengers in the elevator 10, the number of passengers in the elevator 10, etc. The object detection processing unit 23 also detects large baggage other than people. The object detection process using image analysis in the object detection processing unit 23 simply detects the presence or absence of an object of a certain size, such as a person, from the captured image, and this object detection process can be executed with lighter calculations than the person detection process.

[0022] The object detection processing unit 23 includes an agitation detection unit 23a that detects when a passenger is moving significantly (an agitated state) based on passenger detection, and a retention detection unit 23b that detects when the detected passenger is retained in the elevator 10 for a long period of time.

[0023] The image analysis result determination unit 24 receives the image analysis result from the person detection processing unit 22 and the image analysis result from the image analysis result transmission unit 25, and performs an image analysis result determination process to determine the analysis result. The image analysis result transmission unit 25 transmits the determination result in the image analysis result determination unit 24 to the elevator control device 100 and the monitor system 200.

[0024] The monitor system 200 includes a data management unit 201 and a data distribution unit 202 . The data management unit 201 acquires the determination result from the image analysis device 20 together with the captured image. The data distribution unit 202 distributes the acquired determination results to each of the external systems 210 , 220 , and 230 .

[0025] For example, the first external system 210 is a people flow detection system for a building in which an elevator is installed. The data distribution unit 202 supplies data on the number of elevator users (the number of people getting on and off at each floor) from the image analysis results, and data on the age groups of the users, etc., to the first external system 210. This allows the first external system 210, which is a people flow detection system, to detect details of the flow of people using the elevator.

[0026] The second external system 220 is a security system for a building in which the elevator is installed. The data distribution unit 202 supplies detailed data such as a face authentication result from the image analysis result to the second external system 220. This allows the second external system 220, which is a security system, to identify the elevator user and perform security processing such as allowing the car 10 to stop at a security-disabled floor.

[0027] Further, the third external system 230 is a service system that displays advertisements on a display device in the car 10. Then, the data allocation unit 202 supplies data such as the age group of elevator users from the image analysis results to the third external system 230. As a result, the third external system 230, which is a service system that displays advertisements, can determine the age group and gender of elevator users and match the advertisements to be displayed on the display device in the car 10 to the age group and gender of passengers in the car 10.

[0028] [Elevator control device hardware configuration] The elevator control device 100, the image analysis device 20, and the monitor system 200 can each be configured by a computer. FIG. 3 shows an example of a hardware configuration when the elevator control device 100 is configured by a computer. The computer that is the elevator control device 100 is composed of a CPU (Central Processing Unit) 191, a memory 192, a storage 193, an image input / output unit 194, a data input / output unit 195, and an interface (I / F) 196.

[0029] The CPU 191 executes a program stored in the memory 192 or the storage 193. The memory 192 is configured, for example, by a RAM (Random Access Memory), and stores computer programs and calculation result data, and provides the CPU 191 with a work area necessary for each process. In this embodiment, by the CPU 191 executing a program, a door opening state acquisition unit 111, an operation state acquisition unit 112, a door opening state determination unit 113, an operation state determination unit 114, an elevator state management unit 115, and an elevator control processing unit 116 are configured in a work area in the memory 192.

[0030] The image input / output unit 194 inputs the image captured by the camera 12 from the image analysis device 20, and outputs the image to the monitor system 200 or the like. The data input / output unit 195 receives inputs of image analysis results from the image analyzer 20, various data from the operation state detection unit 101 and the door 11, etc. In addition, the data input / output unit 195 outputs switching commands to the image analysis method switching unit 21, elevator states, commands to the operation control unit 102, etc. The interface 196 performs communication processing with other devices via the connected network.

[0031] The image analysis device 20 and the monitor system 200 can also be configured by a computer with a hardware configuration similar to that of the elevator control device 100 shown in FIG. Note that configuring the elevator control device 100, image analysis device 20, and monitor system 200 using a computer is just an example, and for example, a part or all of the processing may be configured using hardware such as a field-programmable gate array (FPGA) or an application specific integrated circuit (ASIC). Also, for image processing, a processor for image processing called a graphics processing unit (GPU) may be used.

[0032] [Analysis processing in image analysis equipment (Example 1: Switching processing based on elevator control commands)] Next, an example of the image analysis method switching process in the image analyzing device 20 in the elevator system of this example will be described. Regarding the image analysis method switching process in the image analyzing device 20, switching may be performed due to various factors, but first, as Example 1, an example of switching processing based on an elevator control command will be described with reference to the flowchart in Fig. 4.

[0033] FIG. 4 is a flowchart showing an example of a process when switching is performed based on an elevator control signal transmitted from the elevator status management unit 115 of the elevator control device 100 to the image analysis method switching unit 21. First, the image analysis method switching unit 21 monitors the elevator control signal transmitted from the elevator control device 100 (step S11). Then, the image analysis method switching unit 21 determines whether or not human detection is enabled by monitoring the elevator control signal (step S12). A state in which human detection is enabled is, for example, a state in which the camera 12 and the image analysis device 20 are operating normally, or a state in which a car call has been made and it is determined that a passenger is using the car, etc.

[0034] When it is determined in step S12 that human detection is enabled (YES in step S12), the image analysis method switching unit 21 selects the image analysis process as the human detection method (step S13) and causes the human detection processing unit 22 to execute detection processing by image analysis. Furthermore, when it is determined in step S12 that human detection is not in an enabled state (NO in step S12), the image analysis method switching unit 21 determines whether or not object detection is in an enabled state (step S14). Whether or not object detection is in an enabled state is determined based on, for example, a control enable / disable signal determined by the elevator or a determination of whether or not it is an elevator service time zone. Alternatively, if the elevator is used but there is no change in the detection state of a person or object for a certain period of time, it is determined that object detection is not enabled. Here, a simple determination may be made based on the setting information of the image analysis device 20.

[0035] When it is determined in step S14 that object detection is enabled (YES in step S14), the image analysis method switching unit 21 selects the image analysis processing as the object detection method (step S15) and causes the object detection processing unit 23 to execute detection processing by image analysis. Moreover, when it is determined in step S14 that the object detection is not effective (NO in step S14), the image analysis method switching unit 21 returns to the monitoring in step S11.

[0036] Then, when image analysis is performed in the person detection processing unit 22 in step S13, and when image analysis is performed in the object detection processing unit 23 in step S15, the image analysis result determination unit 24 acquires the analysis results and performs management processing of the analysis results (step S16). Furthermore, if there is a change in the detection state of a person or object in the management process in step S16, the image analysis result transmission unit 25 transmits the current analysis result to the elevator control device 100 and the monitor system 200 (step S17).

[0037] As shown in the flowchart of Fig. 4, by switching the image analysis method in the image analysis device 20 based on a control signal (control command) from the elevator control device 100, the image analysis device 20 does not need to constantly detect human attributes using the human detection processing unit 22. For example, the image analysis device 20 needs to perform image analysis that places a heavy burden on the human detection processing unit 22 only when the image analysis device 20 is in the passenger transport service. Conversely, when there is little human movement, the image analysis device 20 needs to perform image analysis processing that places a light burden on the object detection processing unit 23. Therefore, the image analysis device 20 can reduce the processing burden of image analysis, reduce the burden on the device, and reduce the power consumption required for image analysis.

[0038] [Analysis processing in an image analyzer (Example 2: Switching processing based on door opening status)] Next, with reference to FIG. 5, another example (example 2) of the image analysis method switching process in the image analyzing device 20 in the elevator system of this example will be described. The example shown in FIG. 5 is a flow chart showing a case where a process of switching the image analysis method is performed based on the door open state of the door 11 of the car 10 and the traveling state.

[0039] First, the image analysis method switching unit 21 monitors the door open state and the operation state based on the signal transmitted from the elevator control device 100 (step S21). Then, the image analysis method switching unit 21 determines whether the door 11 of the car 10 is currently in the door open state or not by monitoring the elevator control signal (step S22). The door open state here may include, for example, a state in which the door 11 is completely open, as well as a state in which the door 11 is in the opening operation and the closing operation. If it is determined in step S22 that the door is open (YES in step S22), the image analysis method switching unit 21 determines whether the elevator car 10 is stopped at any floor (step S23).

[0040] In step S23, if the vehicle is stopped on any floor (YES in step S23), the image analysis method switching unit 21 selects the human detection method for the image analysis process (step S24) and causes the human detection processing unit 22 to execute detection process by image analysis. Also, in step S23, if the vehicle is not stopped on any floor (NO in step S23), the process returns to monitoring in step S21.

[0041] Furthermore, when it is determined in step S22 that the door is not open (NO in step S22), the image analysis method switching unit 21 determines whether or not a predetermined time has elapsed since the elevator car 10 started traveling (step S25). The predetermined time here is a relatively short time, for example, from several seconds to about 10 seconds.

[0042] In step S25, if it is within a predetermined time since the elevator car 10 started moving (YES in step S25), the image analysis method switching unit 21 selects the image analysis processing as the object detection method (step S26) and causes the object detection processing unit 23 to execute detection processing using image analysis. Furthermore, in step S25, if it is not within the predetermined time since the elevator car 10 started traveling (NO in step S25), the image analysis device 20 does not perform image analysis processing, and returns to monitoring in step S21.

[0043] Then, when human image analysis is performed in the human detection processing unit 22 in step S24, and when object image analysis is performed in the object detection processing unit 23 in step S26, if there is a change in the detection state of a person or object, the image analysis result transmission unit 25 transmits the current analysis result as the analysis result to the elevator control device 100 and the monitor system 200 (step S27).

[0044] 5, the image analysis method in the image analysis device 20 is switched based on the door open state signal and the operation state signal from the elevator control device 100, so that when the door is open and there is a high possibility of passenger movement, the image analysis device 20 detects human attributes using the human detection processing unit 22. In other states, the image analysis device 20 performs image analysis processing with a light load using the object detection processing unit 23.

[0045] Therefore, the image analyzing device 20 can reduce the processing load of the image analysis based on the door open state, and can reduce the power consumption required for the image analysis. Furthermore, when a certain amount of time has passed since the start of operation with the doors closed, the image analyzing device 20 stops the image analysis by the object detection processing unit 23 as there is no movement of passengers, and the load on the image analyzing device 20 can be further reduced. Furthermore, by not operating two image analysis algorithms simultaneously, it is possible to realize the algorithm within the limited CPU, GPU, and memory resources of embedded software.

[0046] [Analysis process in image analysis device (Example 3: Another example of switching process based on door open state)] Next, with reference to FIG. 6, another example (example 3) of the image analysis method switching process in the image analyzing device 20 in the elevator system of this example will be described. The example (example 3) shown in Fig. 6 is a flowchart showing a case where image analysis method switching processing is performed based on the door open state and traveling state of the door 11 of the car 10, similarly to the example shown in Fig. 5. In the example (example 3) shown in Fig. 6, both detection by the wildness detection unit 23a and detection by the retention detection unit 23b provided in the object detection processing unit 23 are shown.

[0047] In FIG. 6, the same processes as those in the flowchart of FIG. 5 are given the same step numbers, and duplicated explanations will be omitted. 6, in step S25, it is determined whether or not the time is within a predetermined time (first predetermined time), and if it is not within the first predetermined time, the image analysis method switching unit 21 further determines whether or not the time is within a second predetermined time from the start of travel (step S28). The second predetermined time here is, for example, about one minute to several minutes, and is a time during which it is highly likely that the elevator will continue to travel.

[0048] Then, when it is determined in step S28 that the second predetermined time has not elapsed since the start of travel (YES in step S28), the image analysis method switching unit 21 causes the misbehavior detection unit 23a of the object detection processing unit 23 to detect whether or not a passenger (object) is misbehaving (step S29). Furthermore, when it is determined in step S28 that the second predetermined time has not elapsed since the start of travel (NO in step S28), the image analysis method switching unit 21 detects whether or not a passenger (object) is present using the retention detection unit 23b of the object detection processing unit 23 (step S30).

[0049] Then, in step S27, the analysis result is transmitted when image analysis is performed in the person detection processing unit 22 in step S24, when image analysis is performed in the object detection processing unit 23 in step S26, and when violence is detected in step S29 and when retention is detected in step S30. That is, the image analysis result transmission unit 25 transmits the current analysis result to the elevator control device 100 and the monitor system 200 even when violence is detected in step S29 and when retention is detected in step S30.

[0050] As shown in the flowchart of Fig. 6, by detecting large passenger movements by the movement detection unit 23a during the time when the elevator is thought to be moving, it becomes possible to take measures such as stopping the car 10. Furthermore, by detecting that passengers are still staying in the elevator even after a relatively long time has passed since the elevator started moving, it becomes possible to take measures when passengers are still staying in the elevator.

[0051] [Analysis process in image analysis device (Example 4: Yet another example of switching process based on door open state)] Next, with reference to FIG. 7, another example (example 4) of the image analysis method switching process in the image analyzing device 20 in the elevator system of this example will be described. FIG. 7 is a flowchart showing yet another example of a case where the image analysis method switching process is performed based on the door open state of the door 11 of the car 10. In the example shown in FIG.

[0052] First, the image analysis method switching unit 21 monitors the door open state and the operation state based on a signal transmitted from the elevator control device 100, and determines whether the door 11 of the car 10 is currently in the door open state (step S31). If it is determined in step S31 that the doors are open (YES in step S31), the image analysis method switching unit 21 determines, based on a signal from the elevator control device 100, whether the elevator car 10 has arrived at a floor in response to a hall call (such as a button operation at the hall) and the doors have become open (step S32).

[0053] When it is determined in step S32 that the arrival is due to a platform call (YES in step S32), the image analysis method switching unit 21 selects the person detection method for the image analysis process (step S33), and causes the person detection processing unit 22 to execute detection processing by image analysis. Here, since there is a high possibility that new passengers will appear when a platform call arrives, person detection is performed when the passenger boards the vehicle. Also, in step S32, if it is determined that the arrival was not due to a hall call (NO in step S32), the process returns to the determination in step S31.

[0054] Furthermore, when it is determined in step S31 that the door is not open (NO in step S31), the image analysis method switching unit 21 determines whether or not a destination floor call is currently registered in the car 10 (step S34). The registration of a destination floor call refers to the case where a destination floor button for any floor in the car 10 is pressed and the destination is registered.

[0055] When it is determined in step S32 that a destination floor call has been registered (YES in step S34), the image analysis method switching unit 21 selects the image analysis process as the object detection method (step S35), and causes the object detection processing unit 23 to execute detection processing by image analysis. Here, object detection is performed because the possibility that a new passenger will appear is low and the possibility that a passenger is already present is high when only the destination floor call arrives. Moreover, in step S34, when a destination floor call has not been registered (NO in step S34), analysis is not performed in the image analysis device 20, and the process returns to the monitoring in step S31.

[0056] Then, when image analysis is performed in the human detection processing unit 22 in step S33, and when image analysis is performed in the object detection processing unit 23 in step S35, if there is a change in the detection state of a person or object, the image analysis result transmission unit 25 transmits the current analysis result as the analysis result to the elevator control device 100 and the monitor system 200 (step S36).

[0057] As shown in the flowchart of Fig. 7, the image analysis method in the image analysis device 20 is switched based on the door open state signal from the elevator control device 100, arrival due to a hall call, and whether or not the destination floor is registered, so that the image analysis device 20 can appropriately switch image analysis processing. Specifically, when an arrival due to a hall call occurs, there is a high possibility that a passenger will board the car 10, so the human detection processing unit 22 can appropriately detect the passenger who has boarded by executing passenger attribute detection. On the other hand, in the door closed state (door not open state) when a destination floor call is registered, there is a high possibility that a passenger is in the car 10, so the human detection processing unit 22 can appropriately detect the number of passengers, etc.

[0058] [Analysis process in image analysis device (Example 5: Another example of switching process based on door open state)] Next, with reference to FIG. 8, another example (example 5) of the image analysis method switching process in the image analyzing device 20 in the elevator system of this example will be described. The example (Example 5) shown in FIG. 8 is a flow chart showing a case where the image analysis method switching process is performed based on the door open state of the door 11 of the car 10, similarly to the example shown in FIG. In the example (example 3) shown in FIG. 6, the object detection processing unit 23 performs both detection in the violence detection unit 23a and detection in the stay detection unit 23b. The example shown in FIG. 8 is a flowchart showing a case where the image analysis method switching process is performed based on the door open state of the door 11 of the car 10.

[0059] First, the image analysis method switching unit 21 monitors the door open state and the operation state based on the signal transmitted from the elevator control device 100 (step S41). Then, the image analysis method switching unit 21 determines whether the door 11 of the car 10 is currently in the door open state by monitoring the elevator control signal (step S42). If it is determined in step S42 that the door is open (YES in step S42), the image analysis method switching unit 21 determines whether the elevator car 10 is stopped at any floor (step S43).

[0060] In step S43, if the vehicle is stopped on any floor (YES in step S43), the image analysis method switching unit 21 selects the human detection method as the image analysis process (step S44) and causes the human detection processing unit 22 to execute detection process by image analysis. Also, in step S43, if the vehicle is not stopped on any floor (NO in step S43), the process returns to monitoring in step S41.

[0061] Furthermore, when it is determined in step S42 that the door is not open (NO in step S42), the image analysis method switching unit 21 determines whether or not it is within a first predetermined time since the elevator car 10 started traveling (step S45). The predetermined time here is a relatively short time, for example, from several seconds to about 10 seconds. In step S45, if it is within a predetermined time since the elevator car 10 started moving (YES in step S45), the image analysis method switching unit 21 selects the image analysis processing as the object detection method (step S46) and causes the object detection processing unit 23 to execute detection processing using image analysis. Furthermore, in step S45, if it is not within a predetermined time since the elevator car 10 started traveling (NO in step S45), the image analysis method switching unit 21 further judges whether the elevator car 10 is stopped or not (step S47). If the elevator car 10 is not stopped (NO in step S47), the image analysis method switching unit 21 returns to the monitoring in step S41.

[0062] Furthermore, when it is determined in step S47 that the elevator is stopped (YES in step S47), the image analysis method switching unit 21 determines whether or not the lighting device (not shown) in the car 10 is on (step S48). Here, if the lighting device 13 in the car 10 is on (YES in step S48), the image analysis method switching unit 21 causes the motion detection unit 23a of the object detection processing unit 23 to detect whether or not a passenger (object) is motionless (step S49). Also, if the lighting device 13 in the elevator 10 is not on (NO in step S48), the image analysis method switching unit 21 detects whether or not a passenger (object) is present using the retention detection unit 23b of the object detection processing unit 23 (step S50).

[0063] Then, when image analysis is performed in the person detection processing unit 22 in step S44, when image analysis is performed in the object detection processing unit 23 in step S46, and when violence is detected in step S49 and when lingering is detected in step S50, the image analysis result transmission unit 25 transmits the analysis results (step S51).

[0064] As shown in the flowchart of Figure 8, when the elevator is stopped and the interior lighting device is on, there is a high possibility that a passenger is in the car, so a movement detection process is performed, and when the lighting device is not on, it is unnatural for a passenger to be in the car, so a stagnation detection process is performed, making it possible to appropriately perform movement detection and stagnation detection according to the state of each elevator car 10.

[0065] [Variations] It should be noted that the embodiment examples described so far have been described in detail in order to explain the present invention in an easily understandable manner, and are not necessarily limited to those having all of the configurations described. For example, in the configuration described in Fig. 1 and Fig. 2, the image analysis device 20 is installed in the car 10. However, the image analysis device 20 may be built into the elevator control device 100.

[0066] Furthermore, in the above-described embodiment, the image analysis process is switched by the image analysis method switching unit 21 in the image analysis device 20 receiving a signal from the elevator control device 100. Alternatively, the elevator control device 100 may determine which method of image analysis is better in the processes described in the examples of Figures 4 to 8, and the image analysis method switching unit 21 may execute the switching process by a command based on that determination.

[0067] In addition, the human object detection method may analyze only images during the period when the door sensor is shaded from light during the period from when the door starts to open to when the door is completely closed, thereby further reducing the processing load.

[0068] In addition, the object detection method may target image analysis at the upper body or above, thereby reducing the overlap of people on the in-car camera even when there are many passengers in the car, thereby further improving detection accuracy.

[0069] A model may be trained in advance using a front image of a person as an entry detection model and a rear image of a person as an exit detection model, thereby improving the detection accuracy of the in-car camera.

[0070] Taking into consideration the case where the amount of light shining in from the platform when the doors are open is large, an automatic calibration operation may be performed to automatically move to each floor and open the doors, and the person / object detection threshold may be automatically set according to the amount of light when the doors are opened on each floor. In the case of a platform facing the outside of an apartment building, for example, the amount of light changes depending on the season and time, so by periodically performing a calibration operation during times when the platform is not in use, accuracy according to the environment may be improved.

[0071] The processing load may be further reduced by analyzing only the images taken when the door has been closed and the vehicle has started traveling.

[0072] When performing image analysis while the door is closed, the number of people may be counted only when the amount of movement (amount of change) of people is equal to or less than a predetermined value, thereby improving detection accuracy.

[0073] Detection accuracy may be improved by generating a learning model of a person's back image according to the installation position of the in-car camera.

[0074] Furthermore, in the configuration of FIG. 1, one camera 12 is installed inside the car 10, but multiple cameras may be installed and the analysis of the images captured by each camera may be switched. In addition, in the configuration diagram shown in Figure 2, only the control lines and information lines that are considered necessary for explanation are shown, and not all control lines and information lines in the product are necessarily shown. In reality, it can be considered that almost all components are connected to each other. In addition, the programs that execute the processing in the image analysis device 20 and the elevator control device 100 may be stored in a storage device within the computer, or may be stored in a recording medium such as an external memory, an IC card, an SD card, or an optical disk, and then transferred. [Explanation of symbols]

[0075] 10...car, 11...door, 12...camera, 12a...image management unit, 13...lighting device, 20...image analysis device, 21...image analysis method switching unit, 22...person detection processing unit, 23...object detection processing unit, 23a...violence detection unit, 23b...stay detection unit, 24...image analysis result determination unit, 25...image analysis result transmission unit, 100...elevator control device, 101...operation status detection unit, 102...operation control unit, 111...door opening status acquisition unit, 112...operation Operation state acquisition unit, 113...door opening state determination unit, 114...operation state determination unit, 115...elevator state management unit, 116...elevator control processing unit, 191...CPU, 192...memory, 193...storage, 194...image input / output unit, 195...data input / output unit, 196...interface, 200...monitor system, 201...data management unit, 210...first external system, 220...second external system, 230...third external system

Claims

1. In an elevator system, processing is performed according to images taken by a camera installed inside the elevator car. an object detection processing unit that detects the presence or absence of a person or object in the elevator car based on an analysis of an image captured by the camera; A person detection processing unit that detects attributes of a person in the elevator based on an analysis of an image captured by the camera; an image analysis method switching unit that selectively executes detection in the object detection processing unit and detection in the person detection processing unit; an image analysis result determination unit that determines an image analysis result of either the detection in the object detection processing unit or the detection in the person detection processing unit, Based on the result determined by the image analysis result determination unit, elevator operation control or notification is given to other systems connected to the elevator system. Elevator system.

2. Based on the determination by the operation state determination unit of the elevator control device, the image analysis method switching unit switches between detection by the object detection processing unit and detection by the person detection processing unit.

2. The elevator system of claim 1.

3. when a door open state determination unit of the elevator control device determines that the door of the car is in an open state, the image analysis method switching unit causes the person detection processing unit to detect a person; When the door open state determination unit determines that the door of the car is in a closed state, the image analysis method switching unit causes the object detection processing unit to detect an object.

2. The elevator system of claim 1.

4. Furthermore, the door open state determination unit determines that the door of the car is in the open state when the door of the car is in the open state based on a car call at a landing.

4. The elevator system of claim 3.

5. Further, the door open state determination unit determines that the door of the car is in a closed state when a destination floor is registered in the car.

5. The elevator system of claim 4.

6. An operation control unit that controls the operation of the elevator based on the result determined by the image analysis result determination unit, The door opening time at the stopping floor of the elevator is changed based on the result determined by the image analysis result determination unit.

2. The elevator system of claim 1.

7. The result determined by the image analysis result determination unit is transmitted to a people flow detection system that determines the usage status of the elevator or the usage status of the building in which the elevator is installed.

2. The elevator system of claim 1.

8. The result of the judgment made by the image analysis result judgment unit is transmitted to the security system of the building in which the elevator is installed.

2. The elevator system of claim 1.

9. The object detection processing unit has both or either one of a movement detection unit that detects a movement of an object inside the car that is greater than a specified amount, and a retention detection unit that detects a retention of an object inside the car that is greater than a specified amount.

2. The elevator system of claim 1.

10. An elevator control method for performing processing according to an image captured by a camera installed in a car, An object detection process that detects the presence or absence of a person or object in the elevator car based on an analysis of an image captured by the camera; A person detection process that detects attributes of a person in the elevator based on an analysis of an image captured by the camera; an image analysis method switching process for selectively executing detection by the object detection process and detection by the person detection process; an image analysis result determination process for determining an image analysis result of either the detection in the object detection process or the detection in the person detection process; and an image analysis result transmission process for controlling elevator operation or notifying other systems connected to the elevator system based on the results determined by the image analysis result determination process. Elevator control method.

Images