JP2018118843A - Remote monitoring system for elevator - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a monitoring system for an elevator capable of preventing a notification resulting from a position and a behavior of at least one of a person and an object.SOLUTION: A remote monitoring system 1 includes a control microcomputer 20, a monitoring terminal 3, a camera 210 for photographing inside of a car 21 with time, a processing unit 23 for processing captured images, and a storage unit 221 for storing model patterns of persons. The processing unit 23 performs: first processing for extracting each edge from an image; second processing for applying a model pattern of each person to an edge extracted from the image; and third processing for preventing transmission of abnormality information resulting from a behavior of a person from the control microcomputer 20 to the monitoring terminal 3 when coordinates of the model pattern of the person applied in the second processing are at the first specific coordinates or when a change in coordinates of model patterns in images before and after the image applied with the model pattern of the person in the second processing shows the first specific behavior.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a remote monitoring system for remotely monitoring an elevator.

  Conventionally, a remote monitoring system for monitoring an operation state of an elevator has been provided. The remote monitoring system includes an elevator including a control microcomputer for controlling the operation, and a terminal that is communicably connected to the control microcomputer of the elevator and remotely monitors the operation of the elevator (see Patent Document 1).

  In such a system, when an abnormality or the like occurs in the elevator, the monitoring center terminal (monitoring terminal) notifies the monitoring center monitoring staff of the operation status of the elevator based on a signal from the elevator (control microcomputer). The supervisor recognizes that an abnormality or the like has occurred in the elevator by notifying the monitoring terminal, and requests the technician to go to the elevator.

  However, in the above system, since the monitoring terminal only informs that an abnormality or the like of the elevator has occurred, the monitoring person cannot grasp the cause of the occurrence of the abnormality or the like of the elevator.

  The causes of the elevator notification include human artifacts, such as an act of preventing the door closing operation when loading and unloading luggage. Therefore, when there is a notification due to the above-mentioned artificial factors, the technician who receives a request for seconding from the monitoring staff goes to the site in order to confirm the cause of the abnormality of the elevator. At the time of arrival, the cause of the notification may be resolved, and the dispatch of technicians may be wasted.

JP 2003-212447 A

  Therefore, the present invention provides an elevator monitoring system capable of preventing notification resulting from the position or behavior of at least one of a person and an object.

  An elevator remote monitoring system according to the present invention includes a control microcomputer disposed in an elevator, a monitoring terminal connected to the control microcomputer and capable of receiving abnormality information from the control microcomputer, and the inside of the elevator car over time. An imaging device that captures an image; a processing unit that processes an image captured by the imaging device; and a storage unit that stores model data to be compared with an image captured by the imaging device. , Including at least one human model pattern, and the processing unit extracts a first process for extracting edges from at least two images of the plurality of images captured by the imaging device, and the at least two images. A second process of applying the model pattern of the person to each of the extracted edges of the image for the person, and the second process. When the coordinates of the model pattern of the fitted person are at the first specific coordinates, or when the coordinates of the model pattern of the person in the images before and after the image fitted with the model pattern of the person in the second processing are changed When the first specific behavior is shown, transmission of abnormal information from the control microcomputer to the monitoring terminal and blocking transmission of abnormal information due to the behavior of the person to which the person's model pattern is applied And performing a third process.

  In the elevator remote monitoring system, the position and behavior of a person can be properly recognized by the model pattern of the person. Therefore, when a person is staying at a position where there is a high possibility that abnormal information will be output, and when a person is performing a behavior where there is a high possibility that abnormal information will be output, that is, from the position of the person When the cause of the output of abnormal information is resolved by moving or stopping this behavior by a person, it is caused by the position and behavior of these people (for human causes) Notification (transmission of abnormal information from the control microcomputer to the monitoring terminal) can be prevented.

  In the elevator remote monitoring system, the model data also includes a model pattern of an internal structure of at least one car, and the second processing includes a car at an edge extracted from one of the at least two images. A first step of applying a model pattern of the internal structure of the car to an edge of an image of the internal structure of the car, and an edge other than an edge that overlaps the model pattern of the internal structure of the car among the edges extracted from the one image A second step of applying the model pattern of the person, and an edge other than an edge that overlaps a model pattern of the internal structure of the car in an edge extracted from an image other than the one image of the at least two images, The third step to apply the model pattern of the person applied in the second step Tsu may have a flop.

  According to such a configuration, the human model pattern can be more appropriately applied by applying the human model pattern to edges other than the edge that overlaps the model pattern of the internal structure of the cage among the edges extracted from the captured image. Therefore, the position and behavior of a person can be recognized more appropriately. Thereby, the alerting | reporting resulting from a person's position and behavior can be prevented more appropriately.

  In another elevator remote monitoring system of the present invention, a control microcomputer disposed in an elevator, a monitoring terminal connected to the control microcomputer and capable of receiving abnormality information from the control microcomputer, and the elevator car over time An image pickup apparatus that picks up images, a processing section that processes an image picked up by the image pickup apparatus, and a storage section that stores model data to be compared with an image picked up by the image pickup apparatus. The data includes a model pattern of at least one object, and the processing unit extracts a first process for extracting edges from at least two images out of a plurality of images captured by the imaging device, and the at least two models. A second process of applying a model pattern of the object to each edge of the image of the edge extracted from the image; Changes in the coordinates of the object model pattern in the preceding and succeeding images when the coordinates of the object model pattern applied in theory are at the second specific coordinates, or in the image in which the object model pattern is applied in the second processing. Transmission of abnormality information from the control microcomputer to the monitoring terminal when the second specific behavior is indicated, transmission of abnormality information due to the position or behavior of the object to which the model pattern of the object is applied And performing a third process for preventing the above.

  In the elevator remote monitoring system, the position and behavior of an object can be properly recognized by the model pattern of the object. Therefore, when an object stays at a position where anomaly information is likely to be output, and when an object is performing a behavior that is likely to be anomalous information output, that is, an object from the position. When the cause of the output of abnormal information is resolved by moving or stopping the behavior of an object, a notification (from the control microcomputer to the monitoring terminal) caused by the position or behavior of these objects Transmission of abnormal information).

  In the elevator remote monitoring system, the model data also includes a model pattern of an internal structure of at least one car, and the second processing includes a car at an edge extracted from one of the at least two images. Applying a model pattern of the internal structure of the car to an edge of an image of the internal structure of the car, and an edge other than an edge overlapping the model pattern of the internal structure of the car in an edge extracted from the one image, A second step of applying a model pattern of an object, and an edge other than an edge that overlaps a model pattern of an internal structure of the cage in an edge extracted from an image other than the one image of the at least two images. The third step to apply the model pattern of the fitted object in the step Tsu may have a flop.

  According to such a configuration, the object model pattern can be more appropriately applied by applying the object model pattern to edges other than the edge that overlaps the model pattern of the internal structure of the cage among the edges extracted from the captured image. Because of this, the position and behavior of objects can be recognized more appropriately. Thereby, the alerting | reporting resulting from the position and behavior of an object can be prevented more appropriately.

  In the elevator remote monitoring system, the model data also includes a model pattern of at least one object, and in the second process, the edge of the image with respect to the object among the edges extracted from the at least two images, respectively. In the third process, when the coordinates of the model pattern of the object applied in the second process are at the second specific coordinates, or in the second process, the model pattern of the object is applied in the second process. The abnormality information is transmitted from the control microcomputer to the monitoring terminal when the change in the coordinates of the model pattern of the object in the images before and after the fitted image shows a second specific behavior, and the object The transmission of the abnormal information due to the position or behavior of the object to which the model pattern is applied may be prevented.

  According to this configuration, in addition to being able to properly recognize the position and behavior of a person based on the person's model pattern, it is possible to properly recognize the position and behavior of the thing based on the model pattern of the object. Therefore, in addition to the notification caused by the position and behavior of the person, the notification caused by the position and behavior of the object can be prevented.

  In the elevator remote monitoring system, the car includes a door that opens and closes, and the second specific coordinate is in at least one of an area on the surface of the door and an area on the track of the door. May be.

  According to such a configuration, when the coordinates of the object model pattern overlap with an area where the surface of the door exists (when the object and the door are in contact), and the coordinates of the object model pattern are When it corresponds to at least one of the times when it overlaps (when an object is expected to come into contact with the door), transmission of abnormal information from the control microcomputer to the monitoring terminal is blocked, so that the opening / closing of the door is blocked by the object, etc. Can be prevented (notification when there is a high possibility that the cause of the notification is resolved when the technician arrives at the elevator). As a result, it is possible to prevent the dispatch of technicians from being wasted.

  In the elevator remote monitoring system, the change in the coordinate of the model pattern of the object in the front and back images in the second processing in the second specific behavior is a reciprocating motion in the vertical direction or the horizontal direction. Also good.

  According to such a configuration, when a vertical or horizontal reciprocation of an object occurs, for example, a load carried by a person falls, a person bounces a ball, or a person throws a ball on a wall. Because the transmission of abnormal information from the control microcomputer to the monitoring terminal is blocked, notification caused by such an event (the cause of the notification may be resolved when the technician arrives at the elevator) Notification in the case of high probability) can be prevented. As a result, it is possible to prevent the dispatch of technicians from being wasted.

  In the elevator remote monitoring system, the car includes a door that opens and closes, and the first specific coordinate is in at least one of an area where the surface of the door exists and an area on the track of the door. May be.

  According to this configuration, when the coordinates of the person's model pattern overlap with an area where the surface of the door exists (when the person contacts the door), and the coordinates of the person's model pattern overlap with the area on the orbit of the door. When it falls into at least one of the times (when a person is expected to come into contact with the door), the transmission of abnormal information from the control microcomputer to the monitoring terminal is blocked, resulting in the person's obstruction of the door behavior It is possible to prevent the notification (the notification when the cause of the notification is likely to be eliminated when the technician arrives at the elevator). As a result, it is possible to prevent the dispatch of technicians from being wasted.

  In the elevator remote monitoring system, the change in the coordinate of the human model pattern in the front and back images in the second process in the first specific behavior is a reciprocating motion in the vertical direction or the horizontal direction. Also good.

  According to this configuration, when the person is reciprocating in the vertical direction or the horizontal direction, for example, when the person is swinging the car (for example, when the person is jumping), the control of the abnormality information is performed. Since transmission from the microcomputer to the monitoring terminal is blocked, preventing notifications caused by people jumping (notification when the cause of the notification is likely to be resolved when the technician arrives at the elevator) Can do. As a result, it is possible to prevent the dispatch of technicians from being wasted.

  As mentioned above, according to this invention, the alerting | reporting resulting from the position and behavior of at least one of a person and a thing can be prevented.

FIG. 1 is a block diagram of an elevator remote monitoring system according to an embodiment of the present invention. FIG. 2 is a diagram showing a model pattern of a person used in the elevator remote monitoring system according to the embodiment. FIG. 3 is a diagram showing a model pattern of an object used in the elevator remote monitoring system according to the embodiment. FIG. 4 is a diagram showing a model pattern of an internal structure of a car used in the elevator remote monitoring system according to the embodiment. FIG. 5 is a flowchart showing processing of the elevator remote monitoring system according to the embodiment. FIG. 6 is a flowchart showing processing of the elevator remote monitoring system according to the embodiment. FIG. 7 is a block diagram of an elevator remote monitoring system according to a modification of the present invention.

  Hereinafter, a remote monitoring system for an elevator according to the present invention (hereinafter simply referred to as a remote monitoring system) will be described. The remote monitoring system is a system that remotely monitors the operation status of the elevator and notifies the monitoring center monitor of the operation status of the elevator. In this remote monitoring system, the supervisor can recognize that an abnormality or the like has occurred in the elevator by confirming the notification by the monitoring terminal, and can request the technician to go to the elevator.

  As shown in FIG. 1, the remote monitoring system includes a control microcomputer 20 disposed in the elevator 2, a camera (imaging device) 210 that images the car 21 of the elevator 2, and an image (captured image) captured by the camera 210. , A storage unit 221 that stores model data to be compared with the captured image, and a monitoring terminal 3 connected to the control microcomputer 20.

  In the remote monitoring system 1 of the present embodiment, the image recognition device 22 is connected to the camera 210. The storage unit 221 is provided in the image recognition device 22. The processing unit 23 includes an image recognition processing unit 220 provided in the image recognition device 22 and a microcomputer processing unit 200 provided in the control microcomputer 20.

  The remote monitoring system 1 of this embodiment includes a first server 4 connected to the monitoring terminal 3 and a second server 5 connected to the first server 4. The remote monitoring system 1 according to this embodiment includes a remote monitoring unit 6 connected between the control microcomputer 20 and the monitoring terminal 3. Furthermore, the remote monitoring system 1 of the present embodiment includes an administrator terminal 7 and a technician terminal 8 that can be connected to the first server 4.

  In addition, although the one elevator 2 is shown in FIG. 1, the remote monitoring system 1 of this embodiment is provided with a some elevator. Moreover, although one administrator terminal 7 and technician terminal 8 are shown in FIG. 1, the remote monitoring system 1 of the present embodiment includes a plurality of administrator terminals 7 and a plurality of technician terminals 8.

  The elevator 2 is controlled in addition to the car 21 that moves up and down the hoistway, the control microcomputer 20 that outputs information on the operation status of the elevator 2, and the image recognition device 22 connected between the control microcomputer 20 and the camera 210. A voice recognition device 24, a voice guidance device 25, and a display guidance device 26 are connected to the microcomputer 20, respectively.

  In addition to the camera 210, the car 21 is connected to, for example, a car frame including a ceiling, a wall, and a floor, a door that opens and closes, a floor button, a voice recognition device 24, and the voice in the car 21. A microphone (sound collector) 211 that collects sound, a speaker 212 that is connected to the voice guide device 25 and outputs sound into the car 21, and a character information that is connected to the display guide device 26 and is displayed in the car 21. A display (display unit) 213 and the like are included. In the remote monitoring system 1 of the present embodiment, the car 21 is provided with various sensors that can output information on the operation status of the elevator 2.

  For example, the camera 210 is disposed on a ceiling in the car 21 and images the inside of the car 21 over time. Further, the camera 210 can output captured images (a plurality of captured images) in the car 21 to the image recognition device 22. In the car 21, for example, one monitoring camera is provided to monitor the inside of the car 21. In the remote monitoring system 1 of the present embodiment, this monitoring camera is used as the camera 210.

  Specifically, the camera 210 repeatedly performs shooting in the car 21 intermittently (every predetermined time (for example, every second)). Note that the camera 210 may continuously take images in the car 21.

  The image recognition device 22, the voice recognition device 24, the voice guidance device 25, and the display guidance device 26 are, for example, microcomputers provided outside the car 21. Specifically, the voice recognition device 24 can analyze the voice collected by the microphone 211 and output it to the remote monitoring unit 6 via the control microcomputer 20. The voice guidance device 25 can output voice from the speaker 212 based on information input from the remote monitoring unit 6 via the control microcomputer 20. The display guide device 26 can output, for example, character information from the display 213 based on information input from the remote monitoring unit 6 via the control microcomputer 20.

  The control microcomputer 20 is connected to the monitoring terminal 3 via the communication system R1. For example, the communication system R1 can communicate with the monitoring terminal 3 via the remote monitoring unit 6 and the telephone line in order from the control microcomputer 20, and conversely from the monitoring terminal 3 via the telephone line and the remote monitoring unit 6. The communication system can communicate with the control microcomputer 20.

  The microcomputer processing unit 200 can output, for example, an event code to the remote monitoring unit 6 as information related to the operation status of the elevator 2. The event code output from the microcomputer processing unit 200 is a code associated with the operation status of the elevator 2 itself. Specifically, the event code indicates car position information, car door opening / closing information, motor driving information for raising and lowering the car, power failure, and the like. The event code includes a code indicating the abnormality information of the elevator 2.

  In the remote monitoring system 1 of the present embodiment, the microcomputer processing unit 200 outputs an event code indicating this to the remote monitoring unit 6 when the operation state of the elevator 2 is normal. In addition, when the operation status of the elevator 2 is abnormal, the microcomputer processing unit 200 outputs information indicating this to the image recognition device 22, and the operation status of the elevator 2 is abnormal according to the determination result by the image recognition device 22. Or an event code indicating that the operation status of the elevator 2 is not abnormal.

  For example, the microcomputer processing unit 200 detects whether or not the operation state of the elevator 2 is abnormal based on information output by various sensors provided in the car 21. In the remote monitoring system 1 of the present embodiment, when detecting that the operation state of the elevator 2 is abnormal, the microcomputer processing unit 200 outputs information indicating that the operation state is abnormal to the image recognition device 22.

  Further, the microcomputer processing unit 200 outputs information indicating that the operation state of the elevator 2 is abnormal to the image recognition device 22, and then a person or an object is in contact with the door of the car 21, or When the determination result that a part is reciprocating vertically or horizontally is input from the image recognition processing unit 220, an event code indicating that the operation status of the elevator 2 is not abnormal is output to the remote monitoring unit 6, When the determination result that the person or object is not in contact with the door of the car 21 and the person or object is not reciprocating vertically or horizontally is input from the image recognition processing unit 220, the operation of the elevator 2 An event code indicating that the situation is abnormal is output to the remote monitoring unit 6.

  In the remote monitoring system 1 of the present embodiment, the microcomputer processing unit 200 receives a determination result from the image recognition processing unit 220 that a part of a person (for example, a knee) is reciprocating vertically or horizontally. Then, as an event code indicating that the operation state of the elevator 2 is not abnormal, for example, an event code indicating that there is a possibility of human mischief in the car 21 is output.

  In the remote monitoring system 1 of the present embodiment, when the microcomputer processing unit 200 detects that the operation state of the elevator 2 is not abnormal, the microcomputer processing unit 200 outputs an event code indicating that the operation state is not abnormal to the remote monitoring unit 6. .

  In the remote monitoring system 1 of the present embodiment, the image recognition apparatus 22 uses human and object models as model data to be compared with the image recognition processing unit 220 that performs processing for detecting a person from a captured image and the captured image. A storage unit 221 that stores the model pattern and the model pattern of the internal structure of the car 21 of the elevator 2 is included.

  The storage unit 221 stores the outlines (edges) of the person, the object, and the car 21 as model patterns of the internal structure of the person, the object, and the car 21, respectively. The model pattern of people and things is set as a model pattern common to each elevator, for example. The model pattern of the internal structure of the car is set individually for each elevator, for example. Note that the target objects here are objects such as luggage, animals, and the like.

  For example, the storage unit 221 stores a model pattern of a person having a different silhouette as a model pattern of the person. Specifically, the storage unit 221 stores a model pattern of a person with a general silhouette (general clothes), a model pattern of a person wearing a hat, a model pattern of a person wearing a kimono, and the like.

  More specifically, the general human model pattern stored in the storage unit 221 is a simplified human head, ears, torso, arms, and legs as shown in FIG. Consists of lines (wire frames) connecting the contours. Moreover, the model pattern of the thing memorize | stored in the memory | storage part 221 is comprised with the line which connects each edge | side of a cube, as shown in FIG.

  In the remote monitoring system 1 of the present embodiment, the storage unit 221 stores a model pattern of the internal structure of the car output from the first server 4. More specifically, as shown in FIG. 4, the model pattern of the internal structure of the car stored in the storage unit 221 is a line connecting the outline of the car frame 214 including the ceiling, the wall, and the floor, and the door 215 that opens and closes. And a line connecting the contours of the operation panel 216 including floor buttons and the like.

  In the remote monitoring system 1 of the present embodiment, the storage unit 221 stores information on the first specific coordinates and the first specific behavior as information on the position of the person and the behavior of the person, As the information regarding the behavior of the second, the second specific coordinates and the information regarding the second specific behavior are stored. The coordinates here are coordinates in a three-dimensional space with one point in the car as the origin.

  The first specific coordinate is in at least one of the region on the surface of the door of the car 21 and the region on the door track (the opening / closing track of the door). Similarly, the second specific coordinates are also in at least one of the region on the surface of the door of the car 21 and the region on the door track (the opening / closing track of the door).

  The first specific coordinate is an index for determining whether or not the movement of the door by the person such as the door pressing by the person, the opening of the door, and the collision of the body with the door has occurred. Similarly, the second specific coordinate is an index for determining whether or not the obstacle of the door is obstructed by an object such as an object hitting the door or an obstacle present in the door opening. .

  The first specific behavior corresponds to a change in the coordinates of the person's model pattern when the person reciprocates vertically or horizontally (vertically or horizontally). Similarly, the second specific behavior corresponds to a change in the coordinates of the model pattern of the object when the object reciprocates in the vertical direction or the horizontal direction (vertical direction or horizontal direction). The first specific behavior is, for example, an index for determining whether or not the car 21 is shaken due to a person jumping. The second specific behavior is, for example, an index for determining whether a load that a person was carrying or a swing of the car 21 caused by a person bouncing a ball has occurred.

  When the information indicating that the operation state of the elevator 2 is abnormal is input from the control microcomputer 20, the image recognition processing unit 220 captures images (a plurality of imaging images) in the car 21 captured by the camera 210 over time in a certain period. Image) is acquired, and processing for detecting a person from this captured image is performed. This person detection process (first process and second process) will be described in detail below.

  First, the image recognition processing unit 220 performs a first process of extracting edges from at least two images among a plurality of acquired captured images. Specifically, the image recognition processing unit 220 performs preprocessing such as density correction, distortion correction, and noise removal on each of the plurality of acquired images. The boundary that changes to is extracted as an edge.

  The image recognition processing unit 220 applies the person's model pattern to each of the edges of the image for the person among the edges extracted from the at least two images (a plurality of images acquired in the remote monitoring system 1 of the present embodiment). The second process is performed. The image recognition processing unit 220 applies the size of each model pattern (the model pattern of the internal structure of the cage, the model pattern of the person, and the model pattern of the object) and the field of view of the camera 210 (the camera 210) to the extracted edge. The model pattern is applied to the edge extracted from the captured image by matching the coordinates).

  Specifically, the image recognition processing unit 220 is an image of the internal structure of the car 21 at the edge extracted from one of the plurality of acquired images (hereinafter referred to as the first image) in the second process. The first step of fitting the model pattern of the internal structure of the car 21 to the edge of the car 21 is performed. Next, in the second process, the image recognition processing unit 220 uses the edge extracted from the one image as the edge of the image of the person among the edges other than the edge overlapping the model pattern of the internal structure of the car 21. The second step is to apply a human model pattern.

  Thereafter, in the second process, the image recognition processing unit 220 applies a human model pattern in the second step of the at least two images (a plurality of images acquired in the remote monitoring system 1 of the present embodiment). Of the car 21 at the edge extracted from an image other than the image (image other than the first image of at least two images (a plurality of images acquired in the remote monitoring system 1 of the present embodiment)). A third step is performed in which the model pattern of the person applied in the second step is applied to edges other than the edge that overlaps the structural model pattern.

  Further, in the remote monitoring system 1 of the present embodiment, the image recognition processing unit 220 is extracted from at least two images (a plurality of images acquired in the remote monitoring system 1 of the present embodiment) in the second process. The model pattern of the object is applied to the edge of the image for the object among the edges.

  Specifically, in the remote monitoring system 1 of the present embodiment, the image recognition processing unit 220 overlaps the model pattern of the internal structure of the car 21 among the edges extracted from the first image after the third step. The fourth step of applying an object model pattern to an edge that does not overlap with the human model pattern (the edge where the human model pattern did not apply in the third step) is performed. Further, after the fourth step, the image recognition processing unit 220 extracts from the images other than the first image among at least two images (a plurality of images acquired in the remote monitoring system 1 of the present embodiment). The fifth step of applying the model pattern of the object applied in the fourth step to the edge that does not overlap the model pattern of the internal structure of the car 21 at the edge and does not overlap the human model pattern is performed.

  Hereinafter, the first step to the fifth step in the second process described above will be described in more detail. In the first step, the image recognition processing unit 220 superimposes the model pattern of the internal structure of the car on the edge extracted from the first image, and this overlapped part (the edge overlapping the model pattern of the internal structure of the car). Whether there is a human body feature (for example, whether there is a circular edge that overlaps the human head, whether there is a cylindrical edge that overlaps the torso, and an edge that overlaps the extremities) Or not). When the image recognition processing unit 220 determines that there is a circular edge or the like, it is assumed that the car 21 is manned (a person is detected in the car 21), and the circular edge and the human model pattern The size of the human model pattern and the field of view of the camera 210 are matched so that the portion corresponding to the head overlaps, and the human model pattern is applied to the edge extracted from the first image.

  In the second step, the image recognition processing unit 220 is based on the grounding position of the fitted model pattern of the person in the first image (specifically, the grounding position of the person's model pattern and the floor of the car 21). Recognize the position of a person in the space and the space occupied by the person. In addition, when the contact position of the model pattern of the fitted person in the first image cannot be confirmed, the image recognition processing unit 220 determines the position of the person and the space occupied by the person based on the estimated contact position. recognize. Next, the image recognition processing unit 220 calculates the coordinates of each point from the model pattern of the person applied to the edge extracted from the first image. The coordinates of the human model pattern are, for example, the head apex, both ears, both shoulders, both elbows, the tips of both hands, both hips, both knees, both knees, and the toes of both legs as shown by P1 in FIG. Coordinates.

  In the third step, the image recognition processing unit 220 applies the model pattern of the person applied in the second step to an image other than the first image among the at least two images. More specifically, in the third step, the image recognition processing unit 220 determines the internal structure of the cage 21 of the first image from the edges extracted from the images other than the first image among the plurality of acquired images. Except for the edge that overlaps the model pattern, the model pattern of the person applied in the second step is applied to this edge. Thereafter, the image recognition processing unit 220 calculates the coordinates of each point from the model pattern of a person applied to an image other than the first image.

  In the fourth step, the image recognition processing unit 220 removes the edge that overlaps the model pattern of the internal structure of the car and the edge that overlaps the human model pattern from the edge extracted from the first image, and then remains. The determined edge is determined to be the edge of the image with respect to the object. When the image recognition processing unit 220 determines that there is an edge of an image with respect to the object (when it is determined that there is an object in the car 21), the size of the model pattern of the object is matched with the field of view of the camera 210, An object model pattern is applied to an edge extracted from one image. Further, the image recognition processing unit 220 calculates the coordinates of each point from the model pattern of the object applied to the edge extracted from the first image. The coordinates of each point in the object model pattern are, for example, eight corners of a cube, as indicated by P2 in FIG.

  In the fifth step, the image recognition processing unit 220 applies the model pattern of the object applied in the fourth step to an image other than the first image among the at least two images. More specifically, in the fifth step, the image recognition processing unit 220 determines the internal structure of the car 21 of the first image from the edges extracted from the images other than the first image among the plurality of acquired images. Except for the edge that overlaps the model pattern, and further excludes the edge that overlaps the model pattern of the person in the first image, the model pattern of the person applied in the second step is applied to this edge. Thereafter, the image recognition processing unit 220 calculates the coordinates of each point from the model pattern of the object applied to the image other than the first image.

  When the image recognition processing unit 220 determines that the edge extracted from each image does not have a circular edge or the like, it is assumed that the car 21 is unoccupied and the human model pattern is applied to the edge extracted from each image. Then, the model pattern of the object is applied to the edge extracted from each image. In addition, when the image recognition processing unit 220 determines that the edge extracted from each image has no image edge with respect to the object (when it is determined that there is no object in the car 21), the image recognition processing unit 220 detects the object with respect to the edge extracted from each image. The model pattern fitting process ends.

  Hereinafter, the third process performed after the first process and the second process will be described. The image recognition processing unit 220 has the model pattern coordinates of the person applied in the second process at the first specific coordinates (for example, when the person is in contact with the door of the car 21), or the second When a change in the coordinates of the person's model pattern in the captured images before and after the captured image in which the person's model pattern is applied in the processing shows the first specific behavior (that is, the person swings the car 21) A third process for preventing transmission of abnormality information (for example, an event code indicating that the operation state of the elevator 2 is abnormal) from the control microcomputer 20 to the monitoring terminal 3 is performed. The captured images before and after the captured image in which the human model pattern is applied in the second process are captured images whose imaging times are before and after the captured image in which the human model pattern is applied in the second process. Further, the transmission blocked here is transmission of abnormal information caused by the behavior of the person to which the human model pattern is applied.

  Furthermore, in the remote monitoring system 1 according to the present embodiment, in the third process, the image recognition processing unit 220 has the coordinates of the model pattern of the object applied in the second process at the second specific coordinates (that is, the object). Or when the coordinate of the object model pattern in the captured images before and after the captured image in which the object model pattern is applied in the second processing is the second specific behavior. (I.e., when an object is moving so as to shake the car 21), the abnormal information from the control microcomputer 20 to the monitoring terminal 3 (for example, an event code indicating that the operation status of the elevator 2 is abnormal) ) Transmission. This blocked transmission is the transmission of anomaly information resulting from the behavior of the object to which the object model pattern is applied. The captured images before and after the captured image in which the object model pattern is applied in the second process are captured images whose imaging times are before and after the captured image in which the object model pattern is applied in the second process.

  In addition, when there are three or more captured images before and after the captured image in which the human model pattern is applied in the second processing, “when the coordinates of the human model pattern applied in the second processing are at the first specific coordinates. "In addition to when the coordinates of the human model pattern overlap with the first specific coordinates, the coordinates that the human model pattern coordinates are predicted to reach by movement (coordinates by the flow line prediction) are the first specific coordinates. Including when it overlaps with the coordinates of. In addition, when there are three or more captured images before and after the captured image in which the object model pattern is applied in the second process, “when the coordinate of the model pattern of the object applied in the second process is at the second specific coordinate "Also includes the case where the coordinates of the object model pattern predicted to arrive by movement (coordinates by the flow line prediction) overlap with the second specific coordinates.

  Also, the coordinates of the model pattern of the person or object applied in the second process (the coordinates of each point in the model pattern of the person or object applied to the edge extracted from the first image or an image other than the first image) Is in the first and second specific coordinates when the coordinates of each model pattern overlap the first and second specific coordinates in a certain period.

  As described above, in the second process, the image recognition processing unit 220 applies the model pattern of the person or the object to the image (the first image or the image other than the first image), thereby determining the position of the person or the object. I can grasp. In the second process, the image recognition processing unit 220 applies the model pattern of the person or the object to different images (the first image and the image other than the first image), so that the behavior of the person or the object (the person or the object) Change in the movement of objects). Further, in the remote monitoring system 1 of the present embodiment, the image recognition processing unit 220 uses each of the human and object model patterns for three or more captured images in the captured image in which the object model pattern is applied in the second process. In order to calculate the coordinates of the point, in addition to grasping the behavior of the person and the object by the change in the coordinates of the model pattern of the person and the object calculated in the second process, the behavior of the person and the object can be predicted. It is possible to predict the flow line of objects and objects.

  Furthermore, in the remote monitoring system 1 of the present embodiment, the image recognition processing unit 220 causes the person or object to contact the door of the car 21 or the person or object shakes the car 21 in the third process. When it is determined that the vehicle is moving, a determination result to this effect is output to the microcomputer processing unit 200, whereby abnormality information from the control microcomputer 20 to the monitoring terminal 3 (for example, that the operation status of the elevator 2 is abnormal) Transmission of the event code). If the image recognition processing unit 220 determines in the third process that the person or object is not in contact with the door of the car 21 or the person or object does not move so as to shake the car 21, the microcomputer A determination result to this effect is output to the processing unit 200.

  In the remote monitoring system 1 of the present embodiment, the image recognition processing unit 220 outputs such a determination result also to the first server 4, so that the determination is made from the first server 4 to the administrator terminal 7 and the technician terminal 8. Information about the result can be output.

  For example, the monitoring terminal 3 monitors a large number of elevators installed in a plurality of regions. The monitoring terminal 3 is installed in a monitoring center or the like that is a building away from the target elevator. Furthermore, the control microcomputer 20 can input abnormality information (for example, an event code indicating that the operation state of the elevator 2 is abnormal) to the monitoring terminal 3. In the remote monitoring system 1 of the present embodiment, the monitoring terminal 3 has a display unit that displays information such as an input event code.

  A supervisor (operator) is resident in the monitoring center of this embodiment. The monitor can monitor the operation status of the elevator 2 by confirming the display of the event code or the like on the display unit of the monitoring terminal 3.

  The second server 5 is connected to the first server 4 by, for example, a wired line, and is connected by, for example, P2P (Peer to Peer). The second server 5 has a database 50. The database 50 stores information such as car design values for each elevator. Specifically, the database 50 includes, as design values of the car, the size of the car frame in the car (dimensions in the car), the position of the door in the car, the position of the control panel in the car, and the like. It is stored in association with the elevator machine number). The database 50 can be rewritten. For example, information such as a design value of a car in the elevator can be added to the database 50 when a new elevator is installed.

  In the remote monitoring system 1 of the present embodiment, the database 50 stores information (administrator contact information and technician contact information) regarding the administrator terminal 7 and the technician terminal 8 respectively corresponding to each elevator. For example, since the database 50 outputs the information to the administrator terminal 7 and the technician terminal 8 via the Internet, the IP address of the administrator terminal 7 and the IP address of the technician terminal 8 are used as information regarding the administrator terminal 7 and the technician terminal 8. I remember it. Note that the database 50 may store such information in association with the behavior of a person or an object.

  The first server 4 is connected to the monitoring terminal 3 via the Internet, for example. The first server 4 is connected to the control microcomputer 20 via the Internet, the monitoring terminal 3 and the communication system R1.

  The first server 4 can be connected to the administrator terminal 7 via the communication system R2. The communication system R2 is, for example, a communication system that can communicate from the first server 4 to the administrator terminal 7 via the Internet and can communicate from the administrator terminal 7 to the first server 4 via the Internet. The first server 4 can be connected to the engineer terminal 8 via the communication system R3. The communication system R3 is a communication system that can communicate with the technician terminal 8 from the first server 4 via the Internet and can communicate with the first server 4 from the technician terminal 8 via the Internet, for example. The communication systems R2 and R3 are common in part (Internet).

  The first server 4 includes a server processing unit 40 that creates a model pattern of the internal structure of the car. For example, the server processing unit 40 may execute the database before the first processing by the image recognition processing unit 220 (before the first series of processing from the first processing to the third processing described above is performed for the first time) or when the specification of the car is changed. Based on the information such as the design value of the car stored in 50, a model pattern of the internal structure of the car is created, and the created model pattern is output to the image recognition device 22.

  Specifically, the server processing unit 40 acquires information such as the design value of the car 21 from the database 50 and images the car in an unmanned state (empty car state) from the camera 210 via the image recognition device 22 or the like. Get an image. For example, the server processing unit 40 acquires one captured image in a car in an unattended state.

The server processing unit 40 performs preprocessing such as noise removal on the acquired image in the unmanned car, and extracts edges from the preprocessed image. The server processing unit 40 determines the walls, doors, floors of the car 21 from the edges extracted from the captured image based on the shape characteristics of the walls, doors, floors, and ceiling of the car 21 obtained from the design values of the car 21. And recognizing the position of the ceiling. In addition, the server processing unit 40 identifies the position of the camera in the car 21 based on the positional relationship of the wall of the car 21 recognized from the captured image. Further, the server processing unit 40 re-recognizes the positional relationship of the car wall and the like based on the specified camera position, determines the location overlapping the car frame contour and the door contour, and coordinates of the camera viewpoint (see FIG. calculates the P _C) at 4, to create a model pattern of the internal structure of the car (for example, the floor of the four corners, ceiling corners, door and control panel respective coordinates (P3 in FIG. 4) to calculate these Create a model pattern consisting of lines connecting the coordinates of).

  The server processing unit 40 transmits, for example, the coordinates of the camera viewpoint to the image recognition device 22 as the field of view of the camera in the car, and as the model pattern of the internal structure of the car, for example, the four corners of the floor and the ceiling in the car The coordinates of the four corners, doors, and operation panel are output to the image recognition device 22.

  The server processing unit 40 determines whether the image recognition processing unit 220 determines whether the person or the object is in contact with the door of the car 21 or the person or the object shakes the car 21 in the third process. When the determination result of whether or not) is input from the microcomputer processing unit 200 via the communication system R1, the monitoring terminal 3, and the Internet, the administrator terminal 7 corresponding to the notification source (elevator 2) and the database 50 Information regarding the engineer terminal 8 (contact information of the administrator and contact information of the engineer) is acquired, and for example, the determination result is output as character information to the administrator terminal 7 and output to the engineer terminal 8. The determination result to be output is, for example, a message such as “A person is jumping in the car. There is no problem in the elevator 2 itself”.

  The administrator terminal 7 and the technician terminal 8 are each a portable terminal such as a PDA (Personal Digital Assistant) or a notebook personal computer, or a terminal such as a stationary personal computer. Each of the administrator terminal 7 and the engineer terminal 8 includes an output unit that outputs input information, for example. The output unit is a display when the output information is image information or text information, and is a speaker when the output information is audio information. In the remote monitoring system 1 of this embodiment, the output units of the administrator terminal 7 and the technician terminal 8 are displays, and this display outputs the determination result input from the server processing unit 40.

  The administrator terminal 7 is possessed by, for example, an administrator who manages the elevator 2. The technician terminal 8 is, for example, carried by a technician who goes to the elevator 2 in order to deal with the trouble of the elevator 2 when a request for transfer is received.

  Hereinafter, the flow of processing related to the creation of the model pattern of the internal structure of the car by the server processing unit 40 will be described with reference to the flowchart of FIG.

  The server processing unit 40 performs the database 50 before the first processing by the image recognition processing unit 220 (before the first series of processing from the first processing to the third processing described above is performed for the first time) or when the specification of the car 21 is changed. The design value of the car is acquired (S01). Further, the server processing unit 40 acquires a captured image (captured image in a car in an unattended state (empty car state)) from the camera 210 via the image recognition device 22 or the like (S02). The server processing unit 40 extracts the edge of the internal structure of the car from the acquired captured image (S03).

  Based on the edge extracted from the captured image and the design value of the car acquired from the database 50, the server processing unit 40 specifies the position of the camera 210 in the car and creates a model pattern of the internal structure of the car ( S04).

  Thereafter, the server processing unit 40 outputs information on the position of the camera 210 in the car 21 and information on the model pattern of the internal structure of the car to the image recognition device 22 (image recognition processing unit 220) (S05). Specifically, the server processing unit 40 transmits the coordinates of the camera viewpoint to the image recognition device 22 and, as a model pattern of the internal structure of the car, for example, four corners of the floor, four corners of the ceiling, doors, and A model pattern composed of lines (wire frames) connecting the contours of the operation panels is output to the image recognition device 22. The storage unit 221 stores information regarding the output position of the camera 210 and information regarding the model pattern of the internal structure of the car.

  Next, determination processing by the image recognition processing unit 220 will be described with reference to the flowchart of FIG. Hereinafter, the case where the edge extracted from the image includes the edge of an image of a person or an object will be described.

  When the information indicating that the operation state of the elevator 2 is abnormal is input from the control microcomputer 20 (S10), the image recognition processing unit 220 captures images (a plurality of images) in the car 21 captured over time by the camera 210. (Captured image) is acquired (S11).

  The image recognition processing unit 220 performs a first process of extracting edges from at least two images (a plurality of images acquired in the remote monitoring system 1 of the present embodiment) among a plurality of acquired images (S12). ). Specifically, the image recognition processing unit 220 extracts edges after performing preprocessing such as noise removal on each of the plurality of acquired images.

  The image recognition processing unit 220 applies a human model pattern to the edges extracted from the plurality of images extracted in the first process (S12) (S13).

  Specifically, the image recognition processing unit 220 applies the model pattern of the internal structure of the car 21 to the edge of the image with respect to the internal structure of the car 21 at the edge extracted from the first image among the plurality of acquired images. (First step), obtained by applying a human model pattern to an edge other than the edge that overlaps the model pattern of the internal structure of the car 21 among the edges extracted from this one image (second step) The model pattern of the person applied in the second step is applied to the edge other than the edge overlapping the model pattern of the internal structure of the car 21 in the edge extracted from the image other than the first image among the plurality of images (third Step). Further, the image recognition processing unit 220 calculates the coordinates of each point in the human model pattern from the human model pattern applied to the edges extracted from the plurality of acquired images.

  Next, the image recognition processing unit 220 applies an object model pattern to each of the edges extracted from the plurality of images from which the edges have been extracted in the first process (S12) (S14).

  Specifically, the image recognition processing unit 220 applies the object model pattern to the edge extracted from the first image that does not overlap the model pattern of the internal structure of the car 21 and does not overlap the human model pattern. Is applied (fourth step), and does not overlap with the model pattern of the internal structure of the car 21 at the edge extracted from the image other than the first image among the plurality of acquired images and does not overlap with the model pattern of the person. The model pattern of the object applied in the fourth step is applied to the edge (fifth step). Further, the image recognition processing unit 220 calculates the coordinates of each point in the object model pattern from the object model pattern applied to the edge extracted from the plurality of acquired images.

  The processes of S13 and S14 described above correspond to the second process.

  The image recognition processing unit 220 determines the position of the person and the behavior of the person based on the coordinates of each point in the model pattern of the person applied to the edge extracted from the plurality of acquired images (S15). For example, in a certain period, the image recognition processing unit 220 may apply the coordinates of the model pattern of the person applied in the second process (the person applied to the edge extracted from the first image or an image other than the first image). When the coordinates of each point in the model pattern) are at the first specific coordinates, it is considered that the person and the door are in contact, and it is determined that the elevator 2 has an abnormality caused by the position of the person. .

  In addition, when the change in the coordinates of the person's model pattern in the preceding and succeeding images in the captured image in which the person's model pattern is applied in the second process shows the first specific behavior, the image recognition processing unit 220 It is determined that an operation that shakes the car 21 is performed, and it is determined that an abnormality caused by a person's behavior has occurred in the elevator 2.

  Similar to S15, the image recognition processing unit 220 determines the position of the object and the behavior of the object based on the coordinates of each point in the model pattern of the object applied to the edge extracted from the plurality of acquired images (S16). ). For example, in a certain period, the image recognition processing unit 220 has the coordinates of the model pattern of the object applied in the second process or the third process (the edge extracted from the first image or an image other than the first image). When the coordinates of each point of the model pattern of the object that applies to is at the second specific coordinate, it is considered that the object and the door are in contact, and the elevator 2 has an abnormality caused by the position of the object. judge.

  In addition, the image recognition processing unit 220, when the change in the coordinates of the object model pattern in the preceding and succeeding images in the captured image in which the object model pattern is applied in the second process exhibits the second specific behavior, It is determined that the car 21 is moving so as to be shaken, and it is determined that an abnormality due to the behavior of the object has occurred in the elevator 2.

  The image recognition processing unit 220 outputs the determination results in S15 and S16 to the microcomputer processing unit 200 and the server processing unit 40 (S17). Specifically, the image recognition processing unit 220 includes at least one of an abnormality caused by the position of the person, an abnormality caused by the behavior of the person, an abnormality caused by the position of the object, and an abnormality caused by the behavior of the object. When it is determined that one has occurred, this fact is output to the microcomputer processing unit 200 and the server processing unit 40. In the remote monitoring system 1 of the present embodiment, the image recognition processing unit 220 outputs such a determination result to the server processing unit 40 of the first server 4, so that the administrator terminal 7 The engineer terminal 8 is made to output this determination result.

  When the microcomputer processing unit 200 receives an input indicating that an abnormality caused by a person's position or an abnormality caused by a person's behavior has occurred in the elevator 2 (Yes in S18), an event indicating that there is no abnormality in the elevator 2 The code is output to the monitoring terminal 3 (S19).

  The microcomputer processing unit 200 does not input that the abnormality caused by the position of the person in the elevator 2 or the abnormality caused by the behavior of the person has occurred (No in S18), and is caused by the position of the object in the elevator 2. When an error indicating that there is an abnormality due to the behavior of the object or an abnormality due to the behavior of the object is input (Yes in S20), an event code indicating that there is no abnormality in the elevator 2 is output to the monitoring terminal 3 (S19).

  The microcomputer processing unit 200 does not input that the abnormality caused by the position of the person in the elevator 2 or the abnormality caused by the behavior of the person has occurred (No in S18), and is caused by the position of the object in the elevator 2. When it is not input that an abnormality or abnormality due to the behavior of an object has occurred (No in S20), an event code indicating that there is an abnormality in the elevator 2 is output to the monitoring terminal 3 (S21).

  A series of processes from S15 to S21 corresponds to a third process. When an abnormality caused by the position of a person or an object or an abnormality caused by the behavior of a person or an object occurs in the elevator 2 by the third process, abnormality information (an event code indicating that the elevator 2 is abnormal) is displayed. Output to the monitoring terminal 3 can be prevented.

  In the remote monitoring system 1 described above, an edge is extracted from at least two images captured by the camera 210, and a human model pattern is applied to the extracted edge. Position and behavior can be recognized properly. Therefore, when a person stays at a position where an event code indicating an abnormality of the elevator 2 is likely to be output, and a person is performing a behavior that is likely to output an event code indicating an abnormality. When (when a person moves from that position or when a person stops this behavior, the cause of the output of an event code indicating an abnormality is eliminated) It is possible to prevent notification (due to an artificial cause) (transmission of an event code indicating an abnormality from the control microcomputer 20 to the monitoring terminal 3).

  Since the first specific coordinates are in at least one of the area on the surface of the door of the car 21 and the area on the track of the door, the coordinates of each point in the human model pattern are the first specific coordinates. When the person is in contact with the door, that is, when the person contacts the door and when the person is expected to contact the door (when the person obstructs the door behavior), the notification caused by the obstruction of the door behavior by the person is issued. Can be prevented. As a result, it is possible to prevent the dispatch of technicians from being wasted.

  Since the change in the coordinates of the person's model pattern in the first specific behavior is a reciprocating movement in the vertical direction or the horizontal direction, when the vertical reciprocating movement of the person in the vertical direction or the horizontal direction occurs, that is, When an operation that shakes the car is performed (for example, when a person is jumping), it is possible to prevent notification caused by the person jumping. As a result, it is possible to prevent the dispatch of technicians from being wasted.

  According to the remote monitoring system 1, an edge is extracted from at least two images captured by the camera 210, and a model pattern of the object is applied to the extracted edge. Position and behavior can be recognized properly. Therefore, when an object exists at a position where an event code indicating an abnormality of the elevator 2 is likely to be output, and the object is performing a behavior with a high possibility of outputting an event code indicating an abnormality. Sometimes (when an object moves from its position, or when an object stops this behavior, the cause of the output of an event code indicating an abnormality is eliminated), it depends on the position or behavior of the object Notification (transmission of an event code indicating an abnormality from the control microcomputer 20 to the monitoring terminal 3) can be prevented.

  Since the second specific coordinates are in at least one of the area on the surface of the door of the car 21 and the area on the track of the door, the coordinates of each point of the model pattern of the object are the second specific coordinates. When the object is in contact with the door, i.e. when the object is in contact with the door and when the object is expected to contact the door (when the object impedes the door's behavior) Can be prevented. As a result, it is possible to prevent the dispatch of technicians from being wasted.

  Since the change in the coordinate of the model pattern of the object in the second specific behavior is a reciprocating motion in the vertical direction or the horizontal direction, when a vertical reciprocating motion in the vertical direction or in the horizontal direction occurs, for example, a person Preventing notifications caused by shaking in the basket 21 caused by objects when the carrying baggage falls and bounces, or when a person bounces the ball or a person throws the ball against the wall in the basket 21 Can do. As a result, it is possible to prevent the dispatch of technicians from being wasted.

  The image recognition processing unit 220 determines that, for example, the person swings the car 21 (for example, the person is jumping) when the person is reciprocating vertically or horizontally. Then, this determination result is also output to the server processing unit 40 of the first server 4, and this determination result is output from the server processing unit 40 to the administrator terminal 7 and the technician terminal 8. As a result, the image recognition processing unit 220 can alert the manager or the technician about the mischief of the person in the car 21.

  In the remote monitoring system 1 of the present embodiment, since the processing unit 23 uses a simplified model pattern as a model pattern of a person or an object, the model pattern handled by the processing unit 23 uses a detailed model pattern. Smaller than the case. Therefore, processing is possible even with a microcomputer (control microcomputer 20) that is not very high in processing speed used for controlling the elevator 2, that is, the existing control microcomputer 20 can be installed without installing a microcomputer with high processing speed. Can be used to process.

  In the remote monitoring system 1 of the present embodiment, the image recognition processing unit 220 determines whether the coordinates of the model pattern of a person or an object are at the first and second specific coordinates for a certain period. For example, even if a person or an object temporarily contacts the door, the output of abnormality information (an event code indicating that an abnormality has occurred in the elevator 2) from the control microcomputer 20 to the monitoring terminal 3 is prevented. Absent.

  Furthermore, in the remote monitoring system 1 according to the present embodiment, the image recognition processing unit 220 uses a human for three or more images as images before and after a captured image in which a model pattern of a person or an object is applied in the second process. Since the change in the coordinates of the model pattern of the object or the object is determined, the behavior of the person or the object can be grasped more accurately.

  In the remote monitoring system 1 according to the present embodiment, the server processing unit 40 transmits the determination result (for example, person fallen) by the image recognition processing unit 220 in the third process to the administrator terminal 7. Even when the engineer is staying at a place far from the site (elevator 2), the manager can quickly cope with the fall of the person.

  The remote monitoring system of the present invention is not limited to the above-described embodiment, and it is needless to say that various changes can be made without departing from the gist of the present invention. For example, the configuration of another embodiment can be added to the configuration of a certain embodiment, and a part of the configuration of a certain embodiment can be replaced with the configuration of another embodiment. Furthermore, a part of the configuration of an embodiment can be deleted.

  In the remote monitoring system 1 of the above embodiment, the image recognition processing unit 220 in the second process, the edges other than the edges that overlap the model pattern of the internal structure of the car at the edges extracted from at least two of the captured images. Although both the human model pattern and the object model pattern are applied, only one of the human model pattern or the object model pattern may be applied in the second process.

  Even if processing is performed in this way, if a person's model pattern is applied, transmission of abnormal information due to the position and behavior of the person from the control microcomputer to the monitoring terminal is blocked, and notifications due to these are prevented. be able to. In addition, when a model pattern of an object is applied, it is possible to prevent transmission of abnormal information due to the position and behavior of the object from the control microcomputer to the monitoring terminal, thereby preventing notification due to these.

  Moreover, in the remote monitoring system 1 of the said embodiment, the image recognition process part 220 is a model of the internal structure of a cage | basket | car at the edge extracted from one image of at least two images among the captured images in a 2nd process. After applying the pattern, the human model pattern is applied (first step and second step), and at the edges extracted from images other than one of at least two of the captured images, the internal structure of the cage is Although the model pattern of the person was applied without applying the model pattern (third step), the internal structure of the car was also applied to the edges extracted from images other than one of the two images. The model pattern of the person may be applied after applying the model pattern. That is, when applying the human model pattern to the edges extracted from at least two images of the captured images, the human model pattern may be applied after applying the model pattern of the internal structure of the car each time.

  Furthermore, in the remote monitoring system 1 according to the above-described embodiment, the image recognition processing unit 220 in the second process overlaps the model pattern of the internal structure of the cage among the edges extracted from at least two of the captured images. The human model pattern and the object model pattern were applied to the other edges, but without using the model pattern of the internal structure of the car, the human model pattern and the object were extracted to the edges extracted from the at least two images. The model pattern may be applied.

  For example, the image recognition processing unit 220 determines that an object is in the car based on the difference between at least two of the plurality of images acquired from the camera 210 and the captured image in the unmanned car and the change in these images. It may be detected that some or some object is predicted to get in and out of the car.

  Further, instead of creating a model pattern of the internal structure of the car, the image recognition processing unit 220 extracts an edge from the captured image in the unmanned car acquired from the camera 210, and the storage unit 221 adds to the model data. The edge extracted from the captured image in the unmanned cage may be stored. In this case, the storage unit 221 may not store the model pattern of the internal structure of the car 21.

  In this case, the image recognition processing unit 220 applies the model pattern of the person or object to the difference between the edge extracted from the first image and the edge extracted from the captured image in the unmanned cage. The position and behavior of people and objects can be properly recognized by the model pattern of the machine, so that transmission of abnormal information caused by the person's position and behavior from the control microcomputer to the monitoring terminal can be prevented, and notifications caused by these can be prevented. it can.

  In the remote monitoring system 1 of the above embodiment, in the second process, a process (second step) of applying a human model pattern to the edge extracted from the first image among at least two images from which the edge has been extracted is performed, In addition, a process (third step) of applying a human model pattern to an edge extracted from an image other than the first image among the at least two images is performed, and further, the first image of the at least two images is performed. Applying the object model pattern to the edge extracted from (4th step), and finally applying the object model pattern to the edge extracted from the image other than the first image of the previous two images Although (fifth step) was performed, the order of performing these steps is not limited to this. For example, the fourth step may be performed after the second step, and the third step and the fifth step may be performed in order.

  In addition, as described above, the fitting of the human model pattern and the fitting of the object model pattern to the edges extracted from at least two images from which the edges have been extracted may be performed in order as individual processes, respectively. The fitting of the object model pattern and the object model pattern may be performed simultaneously.

  The processing unit 23 applied the model pattern of the person or the object to the edges extracted from the plurality of captured images acquired in the first process. A model pattern of a person or an object may be applied to edges extracted from at least two images of the captured images. Even in this case, the processing unit 23 can grasp the position and behavior of a person or an object.

  Although the processing unit 23 uses the model pattern set in common for each elevator as the human model pattern and the object model pattern, the processing unit 23 may use a model pattern set individually for each elevator. For example, a model pattern of an individual person may be added according to the silhouette of the user in a building where each elevator is installed.

  The human model pattern is a simplified model of human parts (head and torso, etc.) consisting of lines (wire frames) connecting the outlines, but the human parts are drawn in some detail. It may be a model pattern composed of lines (wire frames) connecting the outlines of things. By using a detailed person model pattern, the position and behavior of the person can be grasped in more detail. For example, when the detailed human model pattern includes a line connecting outlines such as the human pupil, nose, and mouth, the human pupil and nose of the human model pattern with respect to the edge extracted from the captured image. By applying the parts corresponding to the mouth and the mouth, it is possible to more reliably determine whether or not the inside of the car 21 is manned.

  Further, the model pattern of the object is a model pattern composed of lines connecting the sides of the cube, but a more detailed model pattern may be used. The object model pattern may include, for example, a model pattern constituted by lines connecting outlines of bags, strollers, and the like.

  Model data (such as a car interior structure model pattern, a person model pattern, and an object model pattern) other than the storage unit 221 of the image recognition device 22, for example, a place different from the elevator 2 (separated from the elevator 2) It may be stored in a server or the like installed at (location).

  The processing unit 23 includes the image recognition processing unit 220 provided in the image recognition device 22 and the microcomputer processing unit 200 provided in the control microcomputer 20, but not limited thereto, all of the processing units 23 are It may be provided in the control microcomputer 20 or may be provided in the image recognition device 22. Further, at least a part of the processing unit 23 is included in a configuration other than the elevator 2 (for example, a server installed in a place different from the elevator 2 (a place away from the elevator 2) from the elevator 2). May be.

  For example, when the processing unit 23 uses a detailed human model pattern, the processing unit 23 and the storage unit 221 are not connected to the elevator 2 because a higher processing capability is required than when a simplified human model pattern is used. You may provide in the server installed in a different place (a place away from the elevator 2).

  A plurality of cameras 210 may be provided in the car 21. Further, as the camera 210, another camera other than the monitoring camera may be used instead of using the monitoring camera. As such a camera, for example, a 360 ° camera capable of simultaneously photographing the entire periphery may be used.

  In the third process, the image recognition processing unit 220 determines contact between the person and the door of the car 21 based on the coordinates of the person's model pattern, and prevents transmission of output of abnormality information. The positional relationship between the door and the door (the distance between the person and the door), the positional relationship between the person and the wall of the car 21, the positional relationship between the person and the ceiling of the car 21, the positional relationship between the person and the floor of the car 21, and By judging the positional relationship between the person and the operation panel, the transmission of output of abnormality information (for example, an event code indicating that abnormality has occurred in the elevator 2) from the control microcomputer 20 to the monitoring terminal 3 is prevented. Also good. For example, if it is determined that the coordinates of both shoulders and both elbows of the person's model pattern are in contact with the floor of the car 21 for a certain period, it is determined that the person has fallen and the control microcomputer 20 While transmitting the output of abnormal information (for example, an event code indicating that an abnormality has occurred in the elevator 2) to the monitoring terminal 3 may be blocked, an output of information related to a person falling may be transmitted.

  The server processing unit 40 outputs the determination result by the image recognition processing unit 220 to the administrator terminal 7 and the technician terminal 8 as character information, but may output it as voice information.

  Further, the server processing unit 40 may output the determination result by the image recognition processing unit 220 to one of the administrator terminal 7 and the technician terminal 8. For example, by sending only the manager terminal 7 about the possibility of mischief in the car 21, it is possible to alert the manager and to suppress the interruption of the work of technicians corresponding to other elevators.

  The remote monitoring system 1 of the above embodiment includes a plurality of elevators, but may include a single elevator. Moreover, although the remote monitoring system 1 was provided with a plurality of administrator terminals 7 and a plurality of technician terminals 8, it may be provided with a single administrator terminal 7 and a plurality of technician terminals 8. A plurality of administrator terminals 7 and one engineer terminal 8 may be provided.

  Further, the remote monitoring system 1 may include a customer terminal that can be connected to the first server 4. The customer terminal is installed in, for example, a hospital or a nursing home for the elderly. In this case, the server processing unit 40 may output the determination result by the image recognition processing unit 220 to the customer terminal. For example, by transmitting information about a person falling to a customer terminal installed in a hospital or a nursing home for the elderly, a nurse or a caregiver can appropriately respond to the person falling. In order to perform such transmission, the database 50 may store information related to the customer terminal corresponding to the elevator in association with the behavior of a person or an object (for example, contact with the nearest hospital of each elevator). You may store the information related to the person's fall with the other party).

  The communication system R1 includes a telephone line. However, as shown in FIG. 7, the communication system R1 may include the Internet instead of the telephone line. For example, the communication system R1 can communicate with the monitoring terminal 3 sequentially from the control microcomputer 20 via the remote monitoring unit 6 and the Internet, and is controlled from the monitoring terminal 3 via the Internet and the remote monitoring unit 6 in reverse. This is a communication system capable of communicating with the microcomputer 20. In this case, the communication systems R1, R2, and R3 are common in part (Internet).

  Moreover, although communication system R1 included the remote monitoring unit 6, it does not need to include this. In this case, the control microcomputer 20 is connected to the first server 4 via the Internet.

DESCRIPTION OF SYMBOLS 1 ... Remote monitoring system, 2 ... Elevator, 20 ... Control microcomputer, 200 ... Microcomputer process part, 21 ... Car, 210 ... Camera, 211 ... Microphone, 212 ... Speaker, 213 ... Display, 22 ... Image recognition apparatus, 220 ... Image Recognition processing unit, 221 ... storage unit, 23 ... processing unit, 24 ... voice recognition device, 25 ... voice guidance device, 26 ... display guidance device, 3 ... monitoring terminal, 4 ... first server, 40 ... server processing unit, 5 ... second server 50 ... database, 6 ... remote monitoring unit, 7 ... manager terminal, 8 ... technician terminal (technician mobile terminal), R1, R2, R3 ... communication line, P1, P2, P3, P C _... coordinates

Claims (9)

A control microcomputer located in the elevator;
A monitoring terminal connected to the control microcomputer and capable of receiving abnormality information from the control microcomputer;
An imaging device that images the interior of the elevator car over time;
A processing unit for processing an image captured by the imaging device;
A storage unit that stores model data to be compared with an image captured by the imaging device;
The model data includes a model pattern of at least one person,
The processor is
A first process of extracting edges from at least two images out of a plurality of images captured by the imaging device;
A second process of applying the model pattern of the person to each of the edges of the image for the person among the edges extracted from the at least two images;
When the coordinates of the human model pattern applied in the second process are at the first specific coordinates, or in the images before and after the image applied with the human model pattern in the second process When the change in coordinates indicates the first specific behavior, the abnormal information is transmitted from the control microcomputer to the monitoring terminal and the abnormal information caused by the behavior of the person to which the human model pattern is applied. And a third process for blocking transmission,
An elevator remote monitoring system characterized by that. The model data also includes a model pattern of the internal structure of at least one car,
The second process includes
Applying a model pattern of the internal structure of the car to an edge of the image relative to the internal structure of the car at an edge extracted from one of the at least two images;
A second step of applying the person's model pattern to an edge other than the edge that overlaps the model pattern of the internal structure of the cage among the edges extracted from the one image;
The model pattern of the person applied in the second step is applied to an edge other than the edge that overlaps the model pattern of the internal structure of the cage in the edge extracted from the image other than the one image of the at least two images. The elevator remote monitoring system according to claim 1, further comprising a third step. A control microcomputer located in the elevator,
A monitoring terminal connected to the control microcomputer and capable of receiving abnormality information from the control microcomputer;
An imaging device that images the interior of the elevator car over time;
A processing unit for processing an image captured by the imaging device;
A storage unit that stores model data to be compared with an image captured by the imaging device;
The model data includes a model pattern of at least one object,
The processor is
A first process of extracting edges from at least two images out of a plurality of images captured by the imaging device;
A second process of applying a model pattern of the object to an edge of an image for the object among the edges extracted from the at least two images;
When the coordinates of the model pattern of the object applied in the second process are at the second specific coordinates, or the model pattern of the object in the preceding and succeeding images in the image applied with the model pattern of the object in the second process Abnormality caused by the position or behavior of the object to which the model pattern of the object is applied when the change in coordinates indicates a second specific behavior, which is transmission of abnormality information from the control microcomputer to the monitoring terminal Performing a third process for preventing transmission of information,
An elevator remote monitoring system characterized by that. The model data also includes a model pattern of the internal structure of at least one car,
The second process includes
Applying a model pattern of the internal structure of the car to an edge of the image relative to the internal structure of the car at an edge extracted from one of the at least two images;
A second step of applying the model pattern of the object to an edge other than the edge that overlaps the model pattern of the internal structure of the cage among the edges extracted from the one image;
The model pattern of the object applied in the second step is applied to an edge other than an edge that overlaps a model pattern of the internal structure of the cage in an edge extracted from an image other than the one image of the at least two images. The elevator remote monitoring system according to claim 3, further comprising a third step. The model data also includes a model pattern of at least one object,
The processor is
In the second process, a model pattern of the object is applied to each edge of the image for the object among the edges extracted from the at least two images,
In the third process, when the coordinates of the model pattern of the object applied in the second process are at the second specific coordinates, or before and after the image in which the model pattern of the object is applied in the second process When the change in the coordinates of the model pattern of the object at the second time shows a second specific behavior, the abnormal information is transmitted from the control microcomputer to the monitoring terminal, and the model pattern of the object is applied. The elevator remote monitoring system according to claim 1, wherein transmission of the abnormality information due to an object position or behavior is prevented. The car includes a door that opens and closes,
The remote of the elevator according to any one of claims 3 to 5, wherein the second specific coordinates are in at least one of an area where the surface of the door exists and an area on the track of the door. Monitoring system.   The change in the coordinate of the model pattern of the object in the front and back images in the second processing in the second specific behavior is a reciprocating motion in the vertical direction or the horizontal direction. The elevator remote monitoring system according to claim 1. The car includes a door that opens and closes,
6. The elevator according to claim 1, wherein the first specific coordinates are in at least one of an area where the surface of the door exists and an area on the track of the door. Remote monitoring system.   The change in the coordinates of the model pattern of the person in the front and back images in the second processing in the first specific behavior is a reciprocating motion in the vertical direction or the horizontal direction. The elevator remote monitoring system according to claim 8.

Images