Object detection device and elevator system

Abstract

An object detection device capable of easily defining a detection area and an elevator system are provided. The object detection device comprises: an area setting unit that sets a plurality of areas for shape data of an object located below a shape data acquisition device provided at the top of the entrance of an elevator car; a first area detection unit that, by comparing shape data when the entrance door is open with shape data when the entrance door is closed, detects a door area that indicates the position of the door among the plurality of areas in the shape data; a second area detection unit that detects one or more second areas among the plurality of areas in the shape data, based on the position of the door area.a detection area setting unit that sets a detection area in at least one of an area comprising at least one section of the door area and an area comprising at least one section of one or more second areas; and a detection processing unit that detects that a detection target object exists in the detection area based on the shape data.

Description

Technical area

[0001] The present disclosure relates to an object detection device and an elevator system. background

[0002] Patent Literature 1 discloses an object detection device of an elevator system. The object detection device captures an image of the entrance of a car. The object detection device defines a detection area near the door and side post areas in the image. The object detection device can detect a detection target located within the detection area. Citation listPatent literature

[0003] Patent Literature 1: JP 2020-152470 A Summary of the inventionTask of the invention

[0004] However, in the object detection device described in Patent Literature 1, the detection area is set based on information about the installation conditions stored in advance, such as information about the car dimensions and information about the installation angle of a camera. This means that the information about the installation conditions must be stored in advance. Therefore, preparation for setting the detection area is required each time in each elevator system.

[0005] The present disclosure was made to solve the problem described above. An object of the present disclosure is to provide an object detection device capable of easily adjusting a detection area and an elevator system. Means of solving the problem

[0006] An object recognition device according to the present disclosure includes an area setting unit that sets a plurality of areas for shape data of an object present below a shape data acquisition device, wherein the shape data acquisition device is provided at a top of an entrance of a car of an elevator, and the shape data has been acquired by the shape data acquisition device; a first area setting unit that recognizes a door area indicating a position of the door from the plurality of areas in the shape data by comparing shape data when a door of the entrance is opened and shape data when the door of the entrance is closed; a second area recognition unit that recognizes one or more second areas among the plurality of areas in the shape data based on a position of the door area detected by the first area recognition unit;a detection area setting unit that sets a detection area in at least one of an area including at least a portion of the door area detected by the first area detection unit and an area including at least a part of the one or more second areas detected by the second area detection unit; and a detection processing unit that detects that a detection target object exists in the detection area set by the area setting unit based on the shape data acquired by the shape data acquisition device.

[0007] An elevator system according to the present disclosure includes a car, a door provided at an entrance of the car, the object detection device provided at a top of the entrance of the car and including the shape data acquisition device that acquires shape data of an object located thereunder, and a control panel that causes an opening or closing movement of the door to be stopped when it is detected by the object detection device that a detection target object exists in a detection area. Advantageous effects of the invention

[0008] According to the present disclosure, an object recognition device sets a recognition area based on acquired shape data. Therefore, it is possible to easily set the recognition area. Brief description of the drawings [ Fig. 1] Fig. 1 is a diagram showing the outline of a building provided with an elevator system according to a first embodiment. [ Fig. 2] Fig. 2 is a perspective view showing the main components of a car of the elevator system according to the first embodiment. [ Fig. 3] Fig. 3 is a perspective view showing the main components of the car of the elevator system according to the first embodiment. [ Fig. 4] Fig. 4 is a perspective view showing the main components of the car of the elevator system according to the first embodiment. [ Fig. 5] Fig. 5 is a diagram showing a range specified in shape data acquired by a shape data acquisition device of the elevator system according to the first embodiment. [ Fig. 6] Fig. 6 is a diagram showing shape data including a region detected by the elevator system according to the first embodiment. [ Fig. 7] Fig. 7 is a diagram showing shape data including a detection area detected by the elevator system according to the first embodiment. [ Fig. 8] Fig. 8 is a diagram showing shape data including a detection area detected by the elevator system according to the first embodiment. [ Fig. 9] Fig. 9 is a diagram showing shape data for which a recognition process of the elevator system according to the first embodiment is performed. [ Fig. 10] Fig. 10 is a flowchart illustrating a method for setting a detection area of ​​a detection processing device of the elevator system according to the first embodiment. [ Fig. 11] Fig. 11 is a perspective view of the main components of a car of an elevator system according to a second embodiment. [ Fig. 12] Fig. 12 is a diagram showing shape data including a detection area detected by the elevator system according to the second embodiment. [ Fig. 13] Fig. 13 is a diagram showing shape data including a detection area detected by the elevator system according to the second embodiment. [ Fig. 14] Fig. 14 is a perspective view showing the main components of the car of the elevator system in the first and second embodiments. [ Fig. 15] Fig. 15 is a perspective view showing the main components of the car of the elevator system in the first and second embodiments. [ Fig. 16] Fig. 16 is a diagram showing a range defined in shape data acquired by the shape data acquisition device of the elevator system in the first and second embodiments. [ Fig. 17] Fig. 17 is a perspective view showing a hall of the elevator system according to the second embodiment. [ Fig. 18] Fig. 18 is a perspective view of the main components of a car of an elevator system according to a third embodiment. [ Fig. 19] Fig. 19 is a perspective view showing the main components of a car of an elevator system according to a third embodiment. [ Fig. 20] Fig. 20 is a diagram showing a second modification of the elevator system of the third embodiment. [ Fig. 21] Fig. 21 is a diagram showing a third modification of the elevator system of the second embodiment. [ Fig. 22] Fig. 22 is a diagram showing a first example of shape data including a detection area set by an elevator system according to a fourth embodiment. [ Fig. 23] Fig. 23 is a diagram showing a second example of shape data including a detection area set by the elevator system according to the fourth embodiment. [ Fig. 24] Fig. 24 is a diagram showing a relationship between a detection area set by an object detection device of the elevator system according to the fourth embodiment and time. [ Fig. 25] Fig. 25 is a diagram showing a detection area set by an elevator system according to a fifth embodiment. [ Fig. 26] Fig. 26 is a hardware configuration diagram of an object detection device of the elevator systems according to the first to fifth embodiments. Description of the embodiments

[0009] The embodiments of the present disclosure will be described with reference to the accompanying drawings. In each drawing, the same or related parts are designated by the same symbols. Duplicate descriptions of such parts are simplified or omitted. First embodiment

[0010] Fig. 1 is a diagram showing the outline of a building provided with an elevator system in the first embodiment.

[0011] In an elevator system 1 of the Fig. 1, an elevator shaft 2 leads through the floors of a building 3. A machine room 4 is provided directly above the elevator shaft 2. A main rope 5 is wound around a traction machine 6 provided in the machine room 4. A car 7 is suspended in the elevator shaft 2 by the main rope 5.

[0012] A plurality of halls 8 are provided on each floor of building 3. Each of the plurality of halls 8 faces the elevator shaft 2. A plurality of hall doors 9 are provided at the entrances of each of the plurality of halls 8. A plurality of hall door frames 10 are provided at each of the entrances of the plurality of halls 8.

[0013] A car door device 11 is provided at the entrance of the car 7. The car door device 11 is provided with a pair of door panels 12 and a driving machine 13. The pair of door panels 12 are provided at the entrance of the car 7 so as to be perpendicular to a getting in / out direction of the car 7. The getting in / out direction of the car 7 is a direction of getting in or out of the car 7 and belongs to the left and right direction of the side. The two door panels 12 can be opened and closed in a direction perpendicular to the getting in / out direction and in a horizontal direction. The pair of door panels 12 can open and close in opposite directions. The driving machine 13 can open and close the two door panels 12.

[0014] A control panel 14 is provided in the machine room 4. The control panel 14 is electrically connected to the traction machine 6 and the drive machine 13. The control panel 14 can control the entire elevator system 1.

[0015] The elevator system 1 is provided with an object detection device 20 as a safety device. The object detection device 20 is provided, for example, on the car 7. The object detection device 20 can monitor the entrance of the car 7 in order to detect an object located near the entrance.

[0016] During normal operation of the elevator system 1, the traction machine 6 is driven by the control unit 14. The car 7 moves up and down according to the drive of the drive machine 6. When the car 7 reaches the hall 8 of a destination floor, the control panel 14 transmits the command to the drive machine 13 to open the two door leaves 12. The drive machine 13 opens the two door leaves 12 in the door opening / closing direction. The drive machine 13 then closes the pair of door leaves 12 in the opening / closing direction based on an instruction from the control panel 14 to close the pair of door leaves 12.

[0017] The object detection device 20 predetermines a detection area around the pair of door leaves 12. The object detection device 20 can detect that a detection target object is within the detection area before the pair of door leaves 12 is opened or closed. When it detects that a detection target object is within the detection area, the object detection device 20 sends the information that it has detected the detection target object to the control panel 14. When the control panel 14 receives the detection information before the pair of door leaves 12 opens, it stops transmitting a command to open the pair of door leaves 12. When the control panel 14 receives the detection information before the pair of door leaves 12 closes, it stops transmitting a command to close the pair of door leaves 12.Instead of aborting the transmission of the command to open or close the pair of door leaves 12, the control panel 14 can transmit a command to stop the movement to open or close the pair of door leaves 12 to the drive machine 13.

[0018] The configuration of the entrance of the cabin 7 and the object detection device 20 is described below using Fig. 2 described.

[0019] Fig. 2 is a perspective view showing the main components of the car of the elevator system in the first embodiment.

[0020] Fig. 2 shows a sketch of the entrance of the cabin 7. The cabin 7 is provided with a floor 7a, an upper doorpost 7b, and a pair of side doorposts 7c and 7d. The floor 7a is a floor inside the cabin 7. The upper post 7b and the two side posts 7c and 7d are provided inside the cabin 7 as a three-sided frame for the entrance of the cabin 7. The upper post 7b is provided at the upper end of the entrance of the cabin 7. The paired side posts 7c and 7d are provided on both sides of the entrance of the cabin 7. The paired side posts 7c and 7d have side post surfaces adjacent to the respective paired door leaves 12. For example, the side post surface of the side post 7c and the side post surface of the side post 7d are opposite each other. There are free spaces for opening and closing between the paired side posts 7c and 7d and the paired door leaves 12.

[0021] The object recognition device 20 is provided with a shape data acquisition device 21 and a detection processing device 22.

[0022] The shape data acquisition device 21 is provided as a shape data acquisition unit on the upper jamb 7b. For example, the shape data acquisition device 21 is located at the center of the upper post 7b in the horizontal direction. The shape data acquisition device 21 is a sensor for acquiring shape data showing the shape of an object present in a detection target area. In the first embodiment, the shape data acquisition device 21 will acquire the shape of an object as shape data in a two-dimensionally controlled format. Specifically, the shape data acquisition device 21 is a camera for capturing an image or video that is a continuous image. In this case, the shape data acquisition device 21 acquires image data or video data as shape data in a two-dimensional format. The shape data acquisition device 21 may be an infrared camera or the like.act if it can capture two-dimensional shape data.

[0023] The shape data acquisition device 21 is provided to acquire shape data of an object located thereunder. Specifically, the shape data acquisition device 21 is provided on the upper door panel 7b so that it can capture an image of the pair of door panels 12, the floor 7a, and the pair of side posts 7c and 7d. The installation angle and position of the shape data acquisition device 21 can be any angle and position that enable the acquisition of shape data of the pair of door panels 12, the floor 7a, and the pair of side posts 7c and 7d. The shape data acquisition device 21 transmits the acquired shape data to the detection processing device 22.

[0024] The shape data acquisition device 21 acquires shape data and transmits the shape data to the detection processing device 22 at regular intervals. The shape data acquisition device 21 detects, based on information from the Fig. 2, or the control panel 14 that the pair of door leaves 12 is in an open or closed state. The shape data acquisition device 21 acquires shape data when the pair of door leaves 12 is in the open state and transmits the door opening time shape data in which the shape data and a flag indicating the door is open are linked to the detection processing device 22. The shape data acquisition device 21 acquires shape data when the pair of door leaves 12 is in the closed state and transmits shape data when the door is closed to the detection processing device 22, linking the shape data and a flag indicating that the door is closed.

[0025] The shape data acquisition device 21 can detect that the pair of door panels 12 is in an open or closed state based on the acquired shape data. Specifically, the shape data acquisition device 21 acquires a series of shape data that are temporally continuous in a state where the pair of door panels 12 continuously perform an opening or closing movement. The shape data acquisition device 21 automatically extracts the shape data at the time of door opening and the shape data at the time of door closing from the series of shape data. In this case, the shape data acquisition device 21 can transmit the shape data of the door opening time in which the extracted shape data and the door opening time flag are linked, or the shape data of the door closing time in which the extracted shape data and the door closing time flag are linked, to the detection processing device 22.For the process of the shape data acquisition device 21 for extracting the door open time or door closed time shape data, a publicly known technique can be applied.

[0026] The recognition processing device 22 performs a recognition area setting process to specify a recognition area for shape data. The detection processing device 22 performs a recognition process to detect a recognition target object located within a set recognition area. The recognition processing device 22 is provided with a storage unit 23, a receiving unit 24, a range setting unit 25, a first area detection unit 26, a second area detection unit 27, a recognition area setting unit 28, and a recognition processing unit 29.

[0027] The storage unit 23 stores the shape data acquired by the shape data acquisition device 21.

[0028] The receiving unit 24 receives the shape data transmitted from the shape data acquisition device 21 and causes the storage unit 23 to store the shape data.

[0029] When setting the recognition area, the area setting unit 25 sets a plurality of areas for shape data. The area setting unit 25 creates information about the set plurality of areas. At this time, the area setting unit 25 sets an area on the surface in the same dimensional space as the shape data as the plurality of areas. Specifically, for example, the area setting unit 25 sets coordinates in two-dimensional space corresponding to an image corresponding to the shape data stored in the storage unit 23. In this case, the area setting unit 25 expresses the plurality of areas using the coordinates of the two-dimensional space.

[0030] Based on the continuity of the image, that is, the shape data stored in the storage unit 23, the intensity information about the interior of the image, and the like, the area setting unit 25 sets the boundaries of objects in the image using an image processing technique such as edge detection. The area setting unit 25 sets a plurality of areas surrounded by each of the boundaries.

[0031] The area setting unit 25 creates information about the plurality of areas when the door is opened based on the door open time data stored in the storage unit 23. The area setting unit 25 creates information about the plurality of areas when the door is closed based on the door close time data stored in the storage unit 23.

[0032] In the area setting unit, the first area recognition unit 26 recognizes a door area as the first area among the plurality of areas included in the plurality of area information generated by the area setting unit 25. The door area is an area corresponding to the pair of door panels 12 and an area corresponding to the position of the pair of door panels 12 in the shape data. Specifically, the first area recognition unit 26 compares the plurality of areas when the door is opened and the plurality of areas when the door is closed to detect a difference. The first area recognition unit 26 recognizes an area whose difference exceeds a threshold value among the plurality of areas when the door is closed as a door area.For example, the first area detection unit 26 attaches a mark indicating that it is a door area to the detected door area among the plurality of areas.

[0033] In the area setting unit, the second area detection unit 27 detects one or more second areas among the plurality of areas included in the plurality of area information generated by the area setting unit 25. The second area detection unit 27 detects the second areas among the plurality of areas when the door is closed based on a positional relationship with the door area detected by the first area detection unit 26. The second area detection unit 27 attaches a flag indicating that it is a detected second area among the plurality of areas.

[0034] The second areas are a first side post area, a second side post area, and a floor area. The first side post area is an area corresponding to one of the paired side posts 7c and 7d and is an area corresponding to the position of one of the paired side posts 7c and 7d in the shape data. The second side post area is an area corresponding to the other of the paired side posts 7c and 7d and is an area corresponding to the position of the other of the paired side posts 7c and 7d in the shape data. The floor area is an area corresponding to the floor 7a and is an area corresponding to the position of the floor 7a in the shape data.

[0035] When setting the detection area, the area setting unit 28 sets one or more detection areas in the dimensional space corresponding to the shape data for the shape data. For example, when adjusting the shape data in two-dimensional space, the detection area setting unit 28 sets one or more detection areas expressed in two dimensions. Specifically, the detection area setting unit 28 sets areas expressed in two-dimensional coordinates as one or more detection areas in the image acquired by the shape data acquisition device 21.

[0036] When setting the one or more detection areas, the detection area setting unit 28 sets the detection areas based on the position of at least one of the first and second areas. Specifically, for example, the area setting unit 28 sets at least one of an area including at least a part of the door area, an area including at least a part of the first side post area, which is a second area, and an area including at least a part of the second side post area, which is a second area, as the detection area.

[0037] The recognition processing unit 29 performs the recognition process. Specifically, when receiving the shape data from the shape data acquisition device 21, the recognition processing unit 29 applies the recognition area set by the recognition area setting unit 28 in the recognition area setting process to the shape data. The recognition processing unit 29 judges, based on a result of the application to the shape data, whether or not a recognition target object exists within the recognition area in the shape data. If a recognition target object exists within the recognition area, the recognition processing unit 29 recognizes the recognition target object located within the recognition area. In this case, the recognition processing unit 29 transmits the information that it has recognized the recognition target object to the control panel 14.

[0038] The open and closed states of the two door leaves 12 are shown below using the Fig. 3 and Fig. 4 described.

[0039] Fig. 3 and Fig. 4 are perspective views showing the main components of the car of the elevator system in the first embodiment.

[0040] Fig. 3 shows the entrance to the car 7 when the pair of door leaves 12 are in the open state. The open state is a state in which the pair of door leaves 12 are fully open. That is, the time at which the door is opened is the time at which the pair of door leaves 12 are fully open.

[0041] Fig. 4 shows the entrance of the car 7 when the pair of door leaves 12 are in the closed state. The closed state is a state in which the pair of door leaves 12 are completely closed. That is, the time at which the door is closed is the time at which the pair of door leaves 12 are completely closed.

[0042] Next, the plurality of areas set by the area setting unit 25 are Fig. 5 described.

[0043] Fig. 5 is a diagram showing the regions specified in the shape data acquired by the shape data acquisition device of the elevator system in the first embodiment.

[0044] Fig. 5(A) is an image indicating information about a plurality of areas when the door is opened, which was created based on the shape data for the door-open time. Fig. 5(B) is an image indicating information about a plurality of areas when the door is closed, which was created based on the shape data for the door-closed time.

[0045] For example, when setting a plurality of regions for shape data that is a two-dimensional RGB image, the region setting unit 25 sets the regions and boundaries in the image based on the continuity in color space. At this time, the region setting unit 25 recognizes, as continuity in color space, groups, each of which is a group of adjacent pixels with similar colors, among a plurality of pixels constituting the RGB image, and groups each of the groups of pixels. The region setting unit 25 recognizes each of the grouped pixel groups as a region. If there is no continuity between one pixel group and another pixel group, the region setting unit 25 detects and sets a boundary between the two groups. The region setting unit 25 creates information about a plurality of regions based on the detected boundaries or regions.

[0046] For example, when setting a plurality of areas for shape data that is a two-dimensional infrared image, the area setting unit 25 sets the boundaries and areas in the image based on the continuity of the infrared intensity. At this time, the area setting unit 25 recognizes, as the continuity of the infrared intensity, groups, each of which is a group of adjacent pixels with similar intensities, among a plurality of pixels constituting the infrared image, and groups each of the groups of pixels. After that, the area setting unit 25 performs a process similar to that of the RGB image example.

[0047] The first area detection unit 26 compares information about the plurality of areas when the door is opened and information about the plurality of areas when the door is closed, and detects such an area where the difference between the two pieces of information exceeds a threshold value in the information about the plurality of areas when the door is closed. For example, the first area detection unit 26 detects Fig. 5(B) an area A0 where the difference exceeds the threshold. The first area detection unit 26 can detect a plurality of areas A0 where the difference exceeds the threshold.

[0048] Next, the processes performed by the first area detection unit 26 and the second area detection unit 27 are described based on Fig. 6 described.

[0049] Fig. 6 is a diagram showing shape data including a region recognized by the elevator system in the first embodiment.

[0050] As in Fig. As shown in Figure 6, the first area detection unit 26 detects the area A0 as the door area A1 in the information about the plurality of areas when the door is closed. When there are a plurality of areas A0, the first area detection unit 26 integrates the plurality of areas A0 as the door area A1. Furthermore, the first area detection unit 26 may designate areas each belonging to the plurality of areas A0 as a plurality of door areas A1.

[0051] When the shape data is two-dimensionally regulated, the second area recognition unit 27 recognizes a first side post area A2 and a second side post area A3, which are second areas, based on a positional relationship with the door area A1. Specifically, the second area recognition unit 27 recognizes, among the plurality of areas, two areas that are adjacent to the door area A1 in the entry / exit direction and are symmetrical in the door opening / closing direction. At this time, the two areas that are symmetrical in the door opening / closing direction may be areas whose positions are approximately symmetrical relative to the door opening / closing direction. The two areas that are symmetrical in the door opening / closing direction may be areas whose shapes are approximately symmetrical.The second area recognition unit 27 recognizes the two areas as the first side post area A2 and the second side post area A3. The second area recognition unit 27 attaches marks indicating the first side post area A2 and the second side post area A3, respectively, to the areas recognized as the first side post area A2 and the second side post area A3 among the plurality of areas. At this time, either of the two areas may be the first side post area A2 or the second side post area A3. In . Fig. 6, the first side post section A2 belongs to side post 7c. The second side post section A3 belongs to side post 7d.

[0052] Thereafter, the second area recognition unit 27 recognizes a floor area A4, which is a second area, based on a positional relationship between the door area A1, the first side post area A2, and the second side post area A3. Specifically, the second area recognition unit 27 recognizes an area adjacent to the door area A1, the first side post area A2, and the second side post area A3, from among the plurality of areas, as the floor area A4. The second area recognition unit 27 affixes a flag indicating that it is the floor area A4 to the area recognized as the floor area A4 among the plurality of areas.

[0053] Next, the process performed by the detection area setting unit 28 is described based on the Fig. 7 and Fig. 8 described.

[0054] Fig. 7 and Fig. 8 are diagrams showing shape data including a detection area set by the elevator system in the first embodiment.

[0055] Fig. 7 shows a first example of the detection area. The detection area setting unit 28 sets a detection area Ad on each portion of the first side post area A2 and a portion of the second side post area A3. For example, the shape of the detection area Ad is a quadrilateral. The detection area setting unit 28 can set the detection area Ad on the entire first side post area A2 and the entire second side post area A3, respectively.

[0056] Fig. Figure 8 shows a second example of the detection area. In the second example, each detection area Ad has a shape such that a portion near the floor area A4 is larger than that of the first example. The portions of the first side post area A2 and the second side post area A3 located near the floor area are sections that are easy for children to touch.

[0057] The detection area may be within the door area, but is not shown. In this case, the detection area may be set to an area that includes at least a portion of the door area and through which the pair of door leaves 12 passes when opening or closing.

[0058] Furthermore, the detection area may be located in the area of ​​hall 8 that is visible when the two door leaves 12 are opened, although this area is not shown. In this case, the detection area may be defined in an area that includes at least a portion of the door area and has a position associated with hall 8.

[0059] The following describes the process of recognition based on Fig. 9 described.

[0060] Fig. 9 is a diagram showing shape data for which the detection of the elevator system is performed in the first embodiment.

[0061] Fig. 9(A) shows shape data for the case where a detection target object is not detected within a detection area. The storage unit 23 stores the shape data for the case where a detection target object is not detected within the detection area as reference shape data.

[0062] Fig. 9(B) shows shape data when a recognition target is present in the recognition area. When a recognition target is present in the recognition area, the shape data acquisition device 21 acquires shape data of the recognition target, which includes an image of the recognition target. The recognition processing unit 29 compares the shape data of the recognition target and the reference shape data, and calculates the difference between the two pieces of shape data. Specifically, when the pieces of shape data are RGB images, the recognition processing unit 29 calculates the color difference between the recognition target shape data and the reference shape data. That is, the recognition processing unit 29 calculates the color differences between a plurality of pixels constituting the recognition target shape data and a plurality of pixels constituting the reference shape data.The recognition processing unit 29 extracts pixels whose calculated pixel color differences exceed a predetermined threshold. If the extracted amount exceeds a threshold set for the recognition range, the recognition processing unit 29 detects that the difference between the two pieces of shape data exceeds a predetermined threshold. The extracted amount may be an indicator such as the number of pixels whose differences exceed the threshold, or an area formed by the pixels whose differences exceed the threshold. If the difference between the two pieces of shape data exceeds the threshold, the recognition processing unit 29 applies the recognition range to the shape data of the detected target.When a position where the difference between the two pieces of shape data exists is included in the recognition range, the recognition processing unit 29 recognizes the recognition target object located in the recognition range.

[0063] If the pieces of shape data are depth images, the recognition processing unit 29 may calculate the difference between the distances as the difference between the two pieces of shape data. Furthermore, the recognition processing unit 29 may identify an image that detects a detection target not by calculating the difference, but by performing edge processing on the shape data of the detection target. In this case, the recognition processing unit 29 may apply the detection range to the shape data of the detection target in which the image is identified, and detect the detection target if the image exists within the detection range.

[0064] Next, Fig. 10 describes an example of setting the detection area.

[0065] Fig. 10 is a flowchart illustrating the method for setting the detection range of the detection processing apparatus of the elevator system in the first embodiment.

[0066] The recognition processing device 22 starts the process of setting the recognition area at an arbitrary time.

[0067] As in Fig. 10, the detection processing device 22 acquires shape data from the shape data acquisition device 21 in step S01. At this time, the recognition processing device 22 acquires the shape data for the door opening time and the door closing time.

[0068] Thereafter, the recognition processing device 22 sets a plurality of areas in step S02.

[0069] Thereafter, in step S03, the recognition processing device 22 recognizes a door area from the plurality of areas.

[0070] Thereafter, in step S04, the recognition processing device 22 recognizes the first side post area and the second side post area from the plurality of areas as second areas.

[0071] Thereafter, in step S05, the recognition processing device 22 sets a portion of the first side post area and a portion of the second side post area as recognition areas.

[0072] Thereafter, the recognition processing device 22 ends the process of setting the recognition area.

[0073] According to the first embodiment described above, the elevator system 1 is provided with the object detection device 20. The object detection device 20 is provided with the area setting unit 25, the first area detection unit 26, the second area detection unit 27, the detection area setting unit 28, and the detection processing unit 29. The first area detection unit 26 detects a door area by comparing the shape data of the open state and the shape data of the closed state of the pair of door panels 12. The second area detection unit 27 detects second areas based on the door area. The detection area setting unit 28 sets a detection area based on at least one of the first area and the second area.That is, the object detection device 20 can set a detection area without using information about installation conditions, such as dimensional information about the booth 7 and installation angle information, via the shape data acquisition device 21. It is not necessary to input the information about the installation conditions into the object detection device 20 in advance. Therefore, it is possible for the object detection device 20 to easily set a detection area. Furthermore, the object detection device 20 can set a detection area even if the dimensions of the booth 7 differ from those previously assumed.

[0074] Furthermore, the object recognition device 20 uses a two-dimensional image containing two-dimensionally regulated shape data as shape data. The second area recognition unit 27 sets the first side post area and the second side post area based on the positional relationship with the door area. Therefore, the object recognition device 20 can recognize areas belonging to the configuration of the interior of the car 7 from a plurality of areas. At this time, the object recognition device 20 can recognize the second areas through a simple algorithm, for example, an algorithm based on whether areas are adjacent to each other in two-dimensional space. This makes it possible to reduce the computational burden for recognizing the second areas.

[0075] Furthermore, the object detection device 20 defines a detection area to at least one area including at least a part of the first side post area and one area including at least a part of the second side post area. Therefore, an object can be prevented from being drawn into any of the gaps between the side posts 7c and 7d and the door panels 12.

[0076] Furthermore, the object detection device 20 defines an area encompassing at least a portion of the door area, through which the pair of door panels 12 pass during opening or closing, as a detection area. This can prevent an object from becoming trapped between the two door panels 12.

[0077] In addition, the object detection device 20 defines an area that encompasses at least a portion of the door area and includes the position associated with hall 8 as a detection area. Therefore, it is possible to detect an object attempting to enter the car 7 in hall 8.

[0078] The elevator system 1 may be an elevator system without a machine room. The elevator system 1 may be an elevator system in which a plurality of cabins are suspended in the elevator shaft.

[0079] The shape data acquisition device 21 may be provided not on the upper post of the car 7, but on the upper post of the three-sided frame of the hall door frame 10. In this case, the shape data acquisition device 21 may be provided for each of the plurality of halls 8.

[0080] The detection processing device 22 may be provided in the engine room 4. Second embodiment

[0081] Fig. Fig. 11 is a perspective view showing the main components of a car of an elevator system in a second embodiment. Parts identical to or corresponding to those of the first embodiment are denoted by the same symbols. Description of the parts is omitted.

[0082] In the second embodiment, a shape data acquisition device 21 is a sensor for acquiring shape data in a three-dimensionally regulated format. The shape data acquisition device 21 measures points indicating a position where an object exists in three-dimensional space and acquires a three-dimensional point cloud, which is a set of the measured points, as shape data. For example, the shape data acquisition device 21 is a sensor using LiDAR (Light Detection and Ranging) technology, a sensor or stereo camera using ToF (Time of Flight) technology, or the like.

[0083] The information about the points measured by the shape data acquisition device 21 includes information about distances from the shape data acquisition device 21 to the measured points and information about angular positions of the measured points relative to the shape data acquisition device 21. The information about the points measured by the shape data acquisition device 21 may further include measurement intensity information about the measured points at the time the measurement was performed by the shape data acquisition device 21. The data about the measured points is information about the reflection intensities, i.e., the intensities of the waves reflected by the points at the time the measurement was performed by the shape data acquisition device 21.For example, the information about the measured points may include information about three-dimensional spatial coordinates, wherein the origin is the measurement / detection position of the shape data acquisition device 21.

[0084] The shape data acquisition device 21 transmits a three-dimensional point cloud including more than a certain number of measurement points as shape data to a detection processing device 22.

[0085] A region setting unit 25 detects surfaces formed by the three-dimensional point cloud represented by the shape data based on the flatness of the three-dimensional point cloud. At this time, the region setting unit 25 extracts a subset of the point cloud that can be considered to lie on the same plane in the three-dimensional point cloud and labels it as a plane. In this way, the region setting unit 25 separately detects the planes formed by the three-dimensional point cloud as a plurality of planes. The region setting unit 25 sets the detected plurality of planes as a plurality of regions. In this case, each of the plurality of planes is a region of a surface that exists within the three-dimensional space.When the surfaces formed by the three-dimensional point cloud are continuous, the area setting unit 25 can recognize the surfaces formed by the three-dimensional point cloud, such as a plane, a curved surface, and a raised surface, as a plurality of surfaces.

[0086] By comparing the plurality of door-closed areas and the plurality of door-opened areas, a first area detection unit 26 detects a door area from the plurality of areas when the door is closed. Specifically, the first area detection unit 26 detects one or more areas that are not present in the plurality of areas when the door is open but are present in the plurality of areas when the door is closed, and detects the one or more areas as the door area.

[0087] A second area recognition unit 27 recognizes a floor area based on a positional relationship between the plurality of areas and the shape data acquisition device 21. Specifically, the second area recognition unit 27 recognizes an area directly opposite to the shape data acquisition device 21 among the plurality of areas as the floor area.

[0088] The second area detection unit 27 detects a first side post area and a second side post area based on a relationship between the door area and the floor area. Specifically, among the plurality of areas, the second area detection unit 27 detects two areas that are orthogonal to the floor area and adjacent to the door area as the first side post area and the second side post area, respectively.

[0089] Next, a detection area in the second embodiment is determined based on Fig. 12 described.

[0090] Fig. 12 is a diagram showing shape data including a detection area detected by the elevator system in the second embodiment. Fig. 12, the detection processing device 22 is not shown.

[0091] As in Fig. As shown in FIG. 12, a region setting unit 28 sets a three-dimensional space area in the three-dimensional space targeted by the measurement of the shape data acquisition device 21 as one or more detection areas when the shape data is three-dimensionally controlled. For example, the region setting unit 28 sets a three-dimensional space area including at least a part of the first side post area as the detection area Ad. The detection area setting unit 28 sets a three-dimensional space area including at least a part of the second side post area as the detection area Ad.

[0092] The shape of the detection area is arbitrarily determined according to a setting location. Fig. 12, the shape of the detection area Ad is a rectangular parallelepiped. For example, the thickness of the detection area Ad from the first side post area is a few centimeters (cm). For example, the thickness of the detection area Ad from the door area is a few centimeters (cm). Specifically, the thickness of the detection area Ad from the first side post area and the thickness from the door area are about 5 to 10 cm. If the thickness of the detection area Ad of the first side post area or the door area is less than 5 cm, there is a possibility that a target object to be detected cannot be accurately detected. If the thickness of the detection area Ad of the first side post area or the door area is more than 10 cm, there is a possibility that an undetectable object will be detected.

[0093] The shape of the detection area corresponding to the first or second side post area may be a shape whose width in the entry / exit direction is wider at a lower section, although this is not shown. That is, a detection area may be designed so that its volume increases in a section that is easy for children to touch.

[0094] The detection area setting unit 28 can set a three-dimensional space area including the space through which a pair of door leaves 12 passes when opening or closing as the detection area, although it is in Fig. 12 is not shown. At this time, the detection area setting unit 28 sets a three-dimensional space area including the door area as the detection area. In this case, the detection processing device 22 is able to detect a person entering or exiting a car 7 when both door panels 12 are open.

[0095] The detection area setting unit 28 can set a three-dimensional area including a position belonging to a hall 8 in which the car 7 stops as a detection area to be Fig. 12 is not shown. At this time, the detection area setting unit 28 can set a three-dimensional space including the door area as the detection area. In this case, the detection processing device 22 is able to detect a person entering the car 7 in the hall 8 when both door leaves 12 are open.

[0096] The storage unit 23 stores shape data in the case where a detection target object is not present in the detection areas as reference shape data.

[0097] A recognition processing unit 29 compares the shape data of the recognition target acquired by the shape data acquisition device 21 with the shape data of the reference shape and calculates the difference between the two pieces of shape data. Specifically, the recognition processing unit 29 calculates a point cloud of measured points that exist in the shape data of the recognition target but not in the shape data of the reference shape as a difference. The recognition processing unit 29 applies the recognition range to the calculated difference of the point cloud and recognizes that a recognition target object exists in the recognition range when the difference of the point cloud that exists within the recognition range is equal to or greater than a certain threshold.

[0098] Next, Fig. 13 describes another example of the detection area.

[0099] Fig. 13 is a diagram showing shape data including a detection area detected by the elevator system in the second embodiment. Fig. 13, an object recognition device 20 is not shown.

[0100] Fig. Figure 13 does not show the pair of door leaves 12 of a double door type, but a single door leaf 12a. The single door leaf 12a is a door leaf for a cabin. For example, the single door leaf 12a is designed with two panels 12b and 12c. The single door leaf 12a is provided at the entrance of the cabin 7 and is movable in the opening / closing direction.

[0101] A gap exists between the two panels 12b and 12c of the single door leaf 12a. The panel 12b is a panel on the inside of the cabin 7 between the two panels 12b and 12c. A surface 12d of the panel 12b that is perpendicular to the panel 12c is a surface that represents a boundary position between the two panels 12b and 12c. If an object is located near the vertical surface 12d, it may enter the gap between the two panels 12b and 12c when the single door leaf 12a moves in an opening direction d.

[0102] The first area detection unit 26 detects, among a plurality of areas, an area corresponding to the surface 12d of the panel 12b as a door edge area. Specifically, the first area detection unit 26 detects, among the plurality of areas, an area parallel to the entry / exit direction of the car 7 as a door edge area. The first area detection unit 26 attaches a mark indicating that it is a door edge area to the detected door edge area from among the plurality of edges. The door edge area is located within a door area. The first area detection unit 26 can detect the door edge area as the door area. That is, the door area mark and the door edge area mark are attached to the door edge area.

[0103] The area setting unit 28 sets a three-dimensional space area including at least a part of the door edge area as the detection area Ad.

[0104] According to the second embodiment described above, the object recognition device 20 performs recognition of a recognition target object using three-dimensional point cloud data, which is three-dimensionally regularized shape data, as shape data. Therefore, the accuracy of recognizing a recognition target object can be improved compared to using two-dimensional regularized shape data.

[0105] Furthermore, the object recognition device 20 defines at least one of the three-dimensional spatial regions including at least the first side post region and the second side post region as a recognition region. Furthermore, the object recognition device 20 defines a three-dimensional spatial region including a portion of the door region through which the pair of door leaves 12 pass when opening or closing as a recognition region. Furthermore, the object recognition device 20 defines a three-dimensional spatial region including a position belonging to the hall as a recognition region. Therefore, it is possible for the object recognition device 20 to detect an object belonging to each region.

[0106] Furthermore, the object detection device 20 detects a door edge area based on shape data. The object detection device 20 defines a detection area corresponding to the door edge area. Therefore, the object detection device 20 can prevent an object from being detected in the door edge area.

[0107] The recognition processing unit 29 of the object recognition device 20 can perform the recognition process using information about the measurement intensity of an object. If the shape data acquisition device 21 is a device that always measures in a constant measurement direction, such as a stereo camera, the positions of the measured points of a three-dimensional point cloud PC0 of the shape data of the reference shape correspond one-to-one to the positions of the measured points of a three-dimensional point cloud PC1 of the shape data of the recognition target. The recognition processing unit 29 calculates differences between the reflection intensities of the measured points of the point cloud PC0 and the reflection intensities of the measured points of the point cloud PC1.If the difference is equal to or greater than a threshold value for the point cloud PC1 present in the detection area, the detection processing unit 29 detects a detection target. At this time, the detection processing unit 29 also uses information about the distances to the measured points to detect the positions of the measured points. If the shape data acquisition device 21 is a device that performs measurement and slightly changes the measurement direction with each measurement, as is the case with some LiDAR models, the detection processing unit 29 obtains the measured points corresponding to the measured points of the point cloud PC1 from among the plurality of measured points of the point cloud PC0. At this time, the detection processing unit 29 regards the measurement points of the point cloud PC0 with the shortest distances to the measurement points of the point cloud PC1 as the corresponding measurement points.Therefore, the object detection device 20 can improve the accuracy of detecting a detection target object.

[0108] By processing three-dimensionally regulated shape data, the object recognition device 20 can improve the recognition accuracy more than when processing two-dimensionally regulated shape data. The following describes the case of two-dimensionally regulated shape data and the case of three-dimensionally regulated shape data based on the Fig. 14 to 16 are described comparatively.

[0109] Fig. 14 and Fig. 15 are perspective views showing the main components of the car of the elevator system of the first and second embodiments. Fig. 16 is a diagram showing the areas defined for shape data acquired by the shape data acquisition device of the elevator system of the first and second embodiments. Fig. 14 to 16, the detection processing device 22 is not shown.

[0110] Fig. Figure 14 shows a first example of the comparative description.

[0111] Fig. 14(A) shows a case where a detection target exists near the first side post area, which is a detection area. A point P is a point where a straight line passing through the shape data acquisition device 21 and the center of gravity of the detection target intersects the first side post area. The detection target is located at a position including the point P.

[0112] In both the case of processing two-dimensionally regulated shape data and the case of processing three-dimensionally regulated shape data, the object recognition device 20 can recognize that the recognition target object exists in this state in the recognition area. That is, in the case where the shape data acquisition device 21 captures an image that is two-dimensionally regulated shape data, the detection processing device 22 recognizes a recognition target object based on the shape data, for example, as shown in Fig. 9(B).

[0113] Fig. 14(B) shows a case where a detection target exists at a position remote from the first side post area, which is a detection area. The detection target is located on a straight line connecting the shape data acquisition device 21 and the point P. In the case where the shape data acquisition device 21 captures an image that is two-dimensionally regulated shape data, the image in this state is almost the same as that in Fig. 14 (A). In this case, the detection processing device 22 judges that the detection target object exists in the detection area and detects the detection target object. That is, the detection processing device 22 erroneously detects the detection target object.

[0114] When shape data is controlled in three-dimensional format, the detection processing device 22 acquires three-dimensional spatial coordinates indicating a detection target as three-dimensional point cloud data. Therefore, the detection processing device 22 does not detect that a detection target is located within the detection range. This means that the probability of a detection target being misdetected by the detection processing device 22 is low.

[0115] Fig. 15 and Fig. 16 shows a second example of the comparative description.

[0116] Fig. Figure 15 shows a case where a detection target exists at a position remote from the first side post area, which is a detection area. A line segment L indicated by a solid line corresponds to the time at which the light incident on the first side post area from the shape data acquisition device 21 is reflected by the first side post area. Fig. 15 the detection target object is located on the line segment L.

[0117] Fig. 16 is an image acquired by the shape data acquisition device 21 of the first embodiment as two-dimensionally regulated shape data in the Fig. 15 shown condition. As shown in Fig. As shown in Fig. 16, the detected detection target object is not within the detection range. When the reflectance of the surface on which the detection range is set, such as the surface of a side post 7c, is high, an image of the detection target object at the position of point P on the surface in the image is shape data. For example, when the material of the surface is metal or the like, the reflectance of the surface is high. In this case, the detection processing device 22 calculates the difference of the image and judges that the detection target object is present within the detection range. That is, the detection processing device 22 erroneously detects the detection target object.

[0118] When the shape data is regulated in three-dimensional format, the recognition processing device 22 does not judge whether an object exists at point P. That is, the probability that the recognition processing device 22 incorrectly recognizes a recognition target object is low.

[0119] As described above, the object recognition device 20 in the second embodiment has high accuracy in recognizing a detection target object compared to the object recognition device 20 in the first embodiment. That is, by using three-dimensionally controlled shape data, it is possible to prevent false recognition by the object recognition device 20.

[0120] When a detection area is set to encompass the position corresponding to the hall 8, the detection unit 29 can estimate whether or not an object, such as a person, a cart, or a robot, has the intention to enter the car 7. An object having the person's intention to enter the car 7 includes the object controlled or operated to move, with the interior of the car 7 as the target.

[0121] Fig. Fig. 17 is a perspective view showing the hall of the elevator system in the second embodiment. In Fig. 17, the detection processing device 22 is not shown.

[0122] In Fig. 17 defines a detection area Ad in hall 8. The shape data acquisition device 21 acquires the shape data of an object moving in hall 8.

[0123] The recognition processing unit 29 calculates a center of gravity position of the object located in the recognition area Ad based on the shape data acquired by the data acquisition device 21. By continuously detecting the center of gravity position of the object, the recognition processing unit 29 calculates a moving direction of the object. The recognition processing unit 29 recognizes whether the object intends to enter the car 7 based on the moving direction of the object and estimates the person's intention. For example, if it is detected that the object is moving in a moving direction d1 that passes in front of the entrance of the car 7, the recognition processing unit 29 judges that the object does not intend to enter the car 7.For example, if it is detected that the object is moving into the interior of the cabin 7 in a movement direction d2, the detection processing unit 29 judges that the object intends to enter the cabin 7.

[0124] If it is detected that the object intends to enter the car 7, the detection processing unit 29 considers the object as a detection target. That is, in this case, the detection processing unit 29 detects that the detection target is located within the detection range.

[0125] Publicly known techniques can be applied to a method for calculating the center of gravity position of an object, a method for calculating the moving direction of the center of gravity position, and a method for judging whether or not there is a person's intention to board.

[0126] According to a modification of the second embodiment described above, when a detection area is set in a three-dimensional space including the hall 8, the object detection device 20 calculates the center of gravity position of an object within the detection area based on shape data and detects the movement direction of the center of gravity position of the object. The object detection device 20 detects whether or not the object intends to enter the car 7 based on the movement direction of the center of gravity position of the object. If it judges that the object intends to enter, the object detection device 20 detects the object as a detection target. Thus, when an object that does not intend to enter passes in front of the entrance of the car 7, the object can be prevented from being detected as a detection target.Furthermore, an object located in front of the entrance of the booth 7 can be prevented from being detected as a detection target. That is, the object detection device 20 can prevent false detection. Third embodiment

[0127] Fig. 18 and Fig. 19 are perspective views showing the main components of a car of an elevator system in a third embodiment. Parts that are the same as or corresponding to those of the first or second embodiment are designated by the same symbols. A description of the parts is omitted. Fig. 18 and Fig. 19, a detection processing device 22 is not shown.

[0128] In the third embodiment, a shape data acquisition device 21 acquires three-dimensionally controlled shape data similar to the second embodiment. The shape data acquisition device 21 is provided at a position corresponding to the door type of a car 7. In this case, the detection processing device 22 detects the position of the shape data acquisition device 21.

[0129] Fig. 18(A) shows a pair of double-door type door leaves 12. In this case, the shape data acquisition device 21 is provided in the central portion of the upper door post 7b.

[0130] A detection range setting unit 28 calculates a first distance D1 from the shape data acquisition device 21 to a first side post area. The detection range setting unit 28 calculates a second distance D2 from the shape data acquisition device 21 to a second side post area. Fig. In the example shown in Fig. 18(A), the first distance D1 and the second distance D2 are equal. The detection range setting unit 28 compares the first distance D1 and the second distance D2 and judges that the first distance D1 and the second distance D2 are equal. In this case, the detection range setting unit 28 detects that the shape data acquisition device 21 is provided in the central portion of the upper post 7b. The detection processing device 22 can detect that the shape data acquisition device 21 is provided in the central portion of the upper post 7b in each operation.

[0131] Fig. 18(B) shows a single door panel 12a. The single door panel 12a is supported on the first side post portion side. In this case, the shape data acquisition device 21 is provided on the second side of the side post portion, at the center of the upper post 7b in the door panel opening / closing direction. That is, the shape data acquisition device 21 is provided on the opposite side of the side on which the single door panel 12a is supported, relative to the center of the upper post 7b.

[0132] In the Fig. In the example shown in FIG. 18(B), the first distance D1 is greater than the second distance D2. The detection range setting unit 28 compares the first distance D1 and the second distance D2 and judges that the first distance D1 is greater than the second distance D2. In this case, the detection range setting unit 28 recognizes that the shape data acquisition device 21 is provided on the second side post area side of the middle portion of the upper post 7b.

[0133] Although not shown, there may be a case where the single door panel 12a is supported on the second side post portion side. In this case, the second distance D2 is greater than the first distance D1. The detection range setting unit 28 compares the first distance D1 and the second distance D2 and judges that the second distance D2 is greater than the first distance D1. In this case, the detection range setting unit 28 recognizes that the shape data acquisition device 21 is provided on the first side post portion side of the middle portion of the upper post 7b.

[0134] Fig. 19 shows a detection area(s) set according to a position of the shape data detecting device 21.

[0135] In the example of Fig. 19(A), the detection areas Ad are arranged on both the first side post portion side and the second side post portion side.

[0136] In the example of Fig. 19(B), a detection area Ad is set on the first side post portion side, where the first side post portion is a side post portion on the side where the single door panel 12a is supported. That is, when the first distance D1 is greater than the second distance D2, the detection area setting unit 28 sets a three-dimensional space area including at least a portion of the first side post portion as the detection area Ad. In this example, a detection area Ad is not set on the second side post portion side when the single door panel 12a starts to open.

[0137] According to the third embodiment described above, an object recognition device 20 recognizes the door type of the car 7 based on shape data. The object recognition device 20 recognizes an installation position of the shape data position recognition device 21 based on the door type of the car 7. The object recognition device 20 sets a detection area according to the installation position of the shape data acquisition device 21. Therefore, the object recognition device 20 can set a detection area at an appropriate position without requiring advance input of information about the installation conditions of the door type.That is, for example, the object detection device 20 can set a detection area only in an area where an event of an object being pulled into the door panels 12 or the door panel 12a may occur, and it is possible to eliminate the setting of a detection area in an unnecessary area. This can improve detection accuracy.

[0138] When the shape data acquisition device 21 acquires an image with two-dimensionally controlled shape data, as in the first embodiment, the object recognition device 20 can recognize the door type of the car 7. In the following, Fig. 20 and Fig. 21 describes two modifications in which the object recognition device 20 recognizes the door type of the cabin 7 based on an image.

[0139] Fig. 20 is a diagram showing a first modification of the elevator system of the third embodiment. Fig. 21 is a diagram showing a second modification of the elevator system of the second embodiment.

[0140] Fig. 20(A) is a horizontal view of the entrance of the car 7. In the first modification, the shape data acquisition device 21 is provided on the second side post portion side, at the center of the upper post 7b in the door panel opening / closing direction. The shape data acquisition device 21 captures an image that is two-dimensionally controlled shape data. A lens of the shape data acquisition device 21 faces vertically downward. That is, the shape data acquisition device 21 captures an image in a direction parallel to the surfaces of the side posts 7c and 7d.

[0141] Fig. 20(B) is an image captured by the shape data acquisition device 21 when the door is closed. The object recognition device 20 recognizes a door area A1, a first side post area A2, and a second side post area A3 in the image.

[0142] Next, the detection area setting unit 28 of the object detection device 20 calculates a first area of ​​the first side post area A2 and a second area of ​​the second side post area A3 in the image, respectively. Regarding the area of ​​a single image in an image, the greater the distance from the shape data acquisition device 21 to an object indicated by the single image, the smaller the area is. The detection area setting unit 28 detects the door area of ​​the car 7 by comparing the first area and the second area. The detection area setting unit 28 detects an installation position of the shape data acquisition device 21 from the door area of ​​the car 7. The detection area setting unit 28 sets a detection area Ad according to the installation position of the shape data acquisition device 21.

[0143] In this example, the first area is smaller than the second area. When the detection area setting unit 28 detects that the first area is smaller than the second area, the detection area setting unit 28 detects that the door of the car 7 is a single-door type, and detects that the shape data acquisition device 21 is provided on the second side post area A3 adjacent to the center of the upper post 7b in the door panel opening / closing direction. In this case, the area setting unit 28 sets a detection area Ad to an area including at least a part of the first side post area A2.

[0144] When judging that the first area and the second area are the same, the detection area setting unit 28 recognizes that the door of the car 7 is a double door and recognizes that the shape data acquisition device 21 is provided in the central portion of the upper door post 7b. In this case, the detection area setting unit 28 sets the detection areas Ad to a range including at least a part of the first side post area A2 and a range including at least a part of the second side post area A3.

[0145] When the detection area setting unit 28 detects that the second area is smaller than the first area, the detection area setting unit 28 detects that the door of the car 7 is a single-door type and detects that the shape data acquisition device 21 is provided on the first side post area A2 adjacent to the center of the upper post 7b in the door panel opening / closing direction. In this case, the area setting unit 28 sets a detection area Ad to an area including at least part of the second side post area A3.

[0146] According to the first modification of the third embodiment described above, the object recognition device 20 recognizes a position where the shape data position recognition device 21 is provided by comparing the first area of ​​the first side post area and the second area of ​​the second side post area. The object recognition device 20 sets a detection area based on the position where the shape data position recognition device 21 is provided. Therefore, in a case where the shape data acquisition device 21 is not tilted for capturing an image, as in this example, the object recognition device 20 can set a detection area at an appropriate position without requiring advance input of door area information.

[0147] In the second, in Fig. In the modification shown in Fig. 21(A), the shape data acquisition device 21 is provided on the second side post portion side at the center of the upper post 7b in the door panel opening / closing direction. The shape data acquisition device 21 captures an image that is two-dimensionally controlled shape data. A lens of the shape data acquisition device 21 faces diagonally to a vertically downward direction. That is, the shape data acquisition device 21 captures an image in an inclined direction with respect to the direction parallel to the surfaces of the side posts 7c and 7d. For example, when the width of the door of the car 7 is larger than the range that the shape data acquisition device 21 can image, the shape data acquisition device 21 is mounted in an inclined posture.

[0148] Fig. Figure 21(B) is an image captured by the shape data acquisition device 21 with the door closed. The solid-line frame F1 is an image captured during the second modification. A dashed-line frame F2 is an image captured during the first modification. That is, the dashed-line frame F2 is the image captured in Fig. 20(B). The viewing angle of the image in the second modification is more inclined toward the first side post area A2 than the viewing angle of the image in the first modification. In the image of the first modification, indicated by the frame F2, the second area is larger than the first area. On the other hand, in the image of the second modification, indicated by the frame F1, the second area is larger than the first area. Therefore, when the shape data acquisition device 21 is installed in a tilted state, there may be a case where a process similar to that of the first modification is not applicable. In the second modification, the object recognition device 20 performs a process different from that of the first modification.

[0149] In the second modification, the detection area setting unit 28 calculates the first area and the second area similarly to the first modification. The detection area setting unit 28 compares the first area with the second area.

[0150] If it is detected that the first area and the second area are different, the detection area setting unit 28 detects that the door of the car 7 is a single door type. Next, the detection area setting unit 28 detects a direction in which the door opens based on shape data. The detection area setting unit 28 detects a position where the shape data acquisition device 21 is provided based on the direction in which the door opens. The detection area setting unit 28 sets a detection area Ad based on the position where the shape data acquisition device 21 is provided.

[0151] Specifically, when it is detected that the first area and the second area are different, the detection area setting unit 28 acquires from the shape data acquisition device 21 a series of shape data that is continuous in time from the time the door closes to the time the door opens. At this time, it can be detected whether the door is opening or closing based on a signal from the engine 13 or the control panel 14, or it can be detected based on the series of shape data.

[0152] The detection area setting unit 28 adds the shift amounts of the images during door opening based on the shape data of the door area when the door is closed and the series of shape data. The shift amounts are added for each of the plurality of pixels constituting the door area. Specifically, the detection area setting unit 28 compares a plurality of pixels constituting the door area A1 in each of a plurality of images included in the series of shape data and a plurality of pixels constituting the door area A1 in the door-at-time image. The detection area setting unit 28 calculates the differences between the pixels constituting the door area A1 in a specific image and the pixels constituting the door area A1 in the image when the door is closed.The detection range setting unit 28 adds the calculated pixel differences corresponding to the plurality of images and considers the result as the shift amount of the images while the door is open. For example, when a pixel at a position (x, y) is focused, the detection range setting unit 28 calculates a difference between a pixel at the position (x, y) in each of the images included in the series of shape data and a pixel at the position (x, y) in the image when the door is closed. The detection range setting unit 28 calculates the difference corresponding to the pixel at the position (x, y) calculated in each image. The detection range setting unit 28 adds the calculated differences corresponding to the pixels at the position (x, y) corresponding to the plurality of images.The detection area setting unit 28 considers the sum of the differences corresponding to the pixels at the position (x, y) as the displacement amount at the position (x, y) in the image while the door is being opened. The detection area setting unit 28 adds the displacement amounts of pixels at positions encompassing the door area A1.

[0153] The detection area setting unit 28 extracts pixels whose image shift amounts exceed a threshold value when the door is opened as a plurality of extracted pixels. The detection area setting unit 28 calculates the center of gravity positions of the plurality of extracted pixels whose shift amounts exceed the threshold value. The detection area setting unit 28 detects a side post area between the first side post area A2 and the second side post area A3 that is closer to the center of gravity positions.

[0154] In this example, the center of gravity positions are closer to the second side post area A3 than to the first side post area A2, although this is Fig. 21 is not shown. When the detection area setting unit 28 detects that the second side post area A3 is closer to the center of gravity positions, the detection area setting unit 28 detects that the shape data acquisition device 21 is located on the second side post area A3 side, at the center of the upper post 7b. In this case, the area setting unit 28 sets a detection area Ad to an area including at least a part of the first side post area A2.

[0155] When judging that the first side post area A2 is closer to the center of gravity positions, the detection area setting unit 28 detects that the shape data acquisition device 21 is provided in the first side post area A2 on the side of the center of the upper post 7b, although not shown. In this case, the area setting unit 28 sets a detection area Ad to an area including at least a part of the second side post area A3.

[0156] For example, in Fig. 21(B) the door of the booth 7 from the right side to the left side. In an image in which the door of the booth 7 opens from the closed state, on the right side of the page, in a section belonging to the door area A1, there is an image of the floor and the like of the hall 8, which is not shown. On the left side of the sheet, in the section belonging to the door area A1, there is an image of the door area A1. Therefore, the displacement amount of the pixel at each position in the door area A1 increases toward the right side of the page. The number of pixels whose displacement amounts exceed the threshold increases toward the right edge of the page. Therefore, regarding the center of gravity positions of the plurality of extracted pixels, more center of gravity positions are located on the right side of the page than in the center of the page in the door area A1.The detection area setting unit 28 can detect that the shape data acquisition device 21 is provided on the second side post area A3 side.

[0157] When judging that the first area and the second area are the same, the detection area setting unit 28 recognizes that the door of the car 7 is a double-door type, and recognizes that the shape data acquisition device 21 is provided on the central portion of the upper door post 7b, similar to the first modification, although not shown. In this case, the detection area setting unit 28 sets the detection area Ad to an area including at least a part of the first side post area A2 and an area including at least a part of the second side post area A3.

[0158] According to the second modification of the above-described third embodiment, when the first area and the second area are different, the object recognition device 20 calculates the center of gravity positions of a plurality of extracted pixels based on a set of shape data and recognizes a position where the shape data acquisition device 21 is provided based on the center of gravity positions. The object recognition device 20 sets a detection area based on the position where the shape data acquisition device 21 is provided. Therefore, even in a case where the shape data acquisition device 21 is installed tilted to capture an image, the object recognition device 20 can set a detection area Ad at an appropriate position without requiring advance input of information about the installation conditions of the door area. Fourth embodiment

[0159] Fig. 22 is a diagram showing a first example of shape data including a detection area detected by an elevator system in a fourth embodiment. Fig. Fig. 23 is a diagram showing a second example of shape data including a detection area defined by the elevator system in the fourth embodiment. Portions that are the same as or corresponding to those of the first to third embodiments are denoted by the same symbols. The description of the portions is omitted. Fig. 22 and Fig. 23, a detection processing device 22 is not shown.

[0160] In the fourth embodiment, the recognition processing device 22 causes the shape of a recognition area to change according to an opening / closing movement of a pair of door leaves 12.

[0161] Fig. Figure 22 shows a first example of the case of an object recognition device 20 using three-dimensionally controlled shape data. A recognition area setting unit 28 of the recognition processing device 22 sets a recognition area Ad in a three-dimensional space area that includes at least a portion of the door area and through which the pair of door panels 12 pass when opening or closing.

[0162] Fig. 22(A) shows a detection area Ad when the pair of door leaves 12 are in a fully opened state. The detection area Ad is located in a three-dimensional space from one of the two door leaves 12 to the other.

[0163] Fig. 22(B) shows a detection area Ad when the door leaves 12 are closing. The state of the closing pair of door leaves 12 is a state between the fully opened state and the fully closed state. The detection area Ad in this case is defined in the three-dimensional space from one of the two door leaves 12 to the other. The width of the detection area Ad in the opening / closing direction of the door in Fig. 22(B) is smaller than the width of the detection area Ad in the opening / closing direction of the door in Fig. 22(A). That is, the volume of the detection area Ad in Fig. 22(B) is smaller than the volume of the detection area Ad in Fig. 22(A).

[0164] By the object detection device 20 changing the detection area Ad according to a door opening / closing movement, as shown in Fig. 22, it can, for example, prevent the pair of door leaves 12 from being incorrectly recognized as a detection target object.

[0165] In the case of the object recognition device 20 also using two-dimensionally controlled shape data, the object recognition device 20 causes the detection area Ad, which is set to an area including at least a part of the door area and through which the pair of door leaves 12 pass during opening and closing, to change according to the door opening / closing movement, similarly to the case of using three-dimensionally controlled shape data, although not shown.

[0166] Fig. Figure 23 shows a second example of the case where the object recognition device 20 uses three-dimensionally controlled shape data. The recognition area setting unit 28 sets a recognition area Ad in a three-dimensional space area that includes at least part of the door area and includes a position associated with a hall 8.

[0167] Fig. 23(A) shows a detection area Ad when the pair of door leaves 12 are in the fully opened state. The detection area Ad has a radial shape passing between the two door leaves 12, with a shape data acquisition device 21 as its center.

[0168] Fig. 23(B) shows a detection area Ad in the case where the two door leaves 12 are closing. The detection area Ad is shorter than the detection area Ad in FIG. 1, corresponding to the distance between the paired door leaves 12. Fig. 23(A). That is, the volume of the detection area Ad in Fig. 23(B) is smaller than the volume of the detection area Ad in Fig. 23(A).

[0169] When the object detection device 20 also uses two-dimensionally controlled shape data, the object detection device 20 causes the shape of the detection area Ad, which is set to an area including at least a portion of the door area and including the position corresponding to the hall 8, to change according to the opening / closing movement of the door, similarly to the case where three-dimensionally controlled shape data is used, although not shown.

[0170] Next, the change in a detection area is described.

[0171] Fig. 24 is a diagram showing a relationship between a detection area set by the object detection device of the elevator system in the fourth embodiment and time.

[0172] Fig. Figure 24 shows graphs, each of which indicates a relationship between time t and the size of the detection area. The horizontal axis of the graphs indicates time.

[0173] The graph labeled (a) indicates a relationship between time t and the opening / closing amount of the pair of door leaves 12. The pair of door leaves 12 are in the fully closed state from time ta0 to ta1. At this time, the opening / closing amount is 0. At time ta1, the pair of door leaves 12 starts an opening movement and is in the fully opened state at time ta2. At this time, the opening / closing amount increases from time ta1 and reaches the largest value at time ta2. From time ta2 to time ta3, the pair of door leaves 12 is in the fully opened state. At time ta3, the pair of door leaves 12 starts a closing movement and is in the fully closed state at time ta4. At this time, the opening / closing amount decreases from time ta3 and becomes 0 at time ta4.

[0174] The graph labeled (b) indicates a relationship between time t and the size of a detection area for detecting retraction. In the case of shape data controlled two-dimensionally as in the first embodiment, the detection area for detecting retraction is a detection area set to an area including at least a part of a first side post area or an area including at least a part of a second side post area. In this case, the size of the detection area corresponds to the area of ​​the detection area.In the case of shape data controlled three-dimensionally as in the second embodiment, the detection area for detecting the retraction is a detection area located in a three-dimensional space area including at least a portion of the first side post area or a three-dimensional space area including at least a portion of the second side post area. In this case, the size of the detection area is the volume of the detection area.

[0175] The size of the detection area for detecting the pull-in becomes the largest before and after the start of the opening movement of the door leaf pair 12, and becomes 0 at times other than this time. When the detection area size is 0, it can be considered that no detection area is set. Specifically, at time tb1 before time ta1, the size of the detection area for detecting the pull-in becomes the maximum value of 0. After time ta1, the size of the detection area for detecting the pull-in becomes the maximum value of 0 at time tb2.

[0176] The graph labeled (c) indicates a relationship between time t and the size of the detection area for detecting door jamming. When the shape data is controlled two-dimensionally as in the first embodiment, the detection area for detecting door jamming is a detection area set to a range in which the pair of door leaves 12 open and close. In this case, the size of the detection area corresponds to the area of ​​the detection area. When the shape data is controlled three-dimensionally as in the second embodiment, the detection area for detecting door jamming is a detection area located in a three-dimensional space in which the pair of door leaves 12 open and close. In this case, the size of the detection area is the volume of the detection area.

[0177] The size of the door jamming detection area is 0 until the pair of door leaves 12 is fully opened. At time tc1, between time ta2 and time ta3, the size of the door jamming detection area reaches the maximum value of 0. After that, at time tc2, which is the same as time ta3, the size of the door jamming detection area begins to decrease from the maximum value as the opening / closing amount of the doors decreases. At time tc3 immediately before time ta4, when the pair of door leaves 12 is fully closed, the size of the door jamming detection area becomes 0.

[0178] The graph labeled (d) indicates the relationship between time t and the size of the detection area for detecting an object attempting to board. In this case, the boarding object is a detection target object that has the human intention of entering the car 7 from the hall 8. When the shape data is controlled two-dimensionally as in the first embodiment, the detection area for detecting an object attempting to board is a detection area including an area belonging to the hall 8. In this case, the size of the detection area corresponds to the area of ​​the detection area. When the shape data is controlled three-dimensionally as in the second embodiment, the detection area for detecting an object attempting to board is a detection area located in a three-dimensional space including the position belonging to the hall 8.In this case, the size of the detection area is the volume of the detection area.

[0179] The size of the detection area for detecting an intruding object is 0 until the pair of door panels 12 are fully opened. At time td1, between time ta2 and time ta3, the size of the detection area for detecting an intruding object reaches the maximum value of 0. After that, at time td2, which coincides with time ta3, the size of the detection area for detecting an intruding object begins to decrease from the maximum value as the opening / closing amount of the doors decreases. At time td3 immediately before time ta4, when the pair of door panels 12 are fully closed, the size of the detection area for detecting an intruding object becomes 0.

[0180] According to the fourth embodiment described above, the object detection device 20 causes the shape of a detection area to change according to the movement of the opening or closing pair of door panels 12. Therefore, it is possible for the object detection device 20 to set only an area required for detection as the detection area. Furthermore, it can prevent the door panels 12 from being mistakenly detected as a detection target when the door panels 12 or the like move into the detection area. Fifth embodiment

[0181] Fig. Fig. 25 is a diagram showing a detection area defined by an elevator system in a fifth embodiment. Portions that are the same as or corresponding to those of the first to fourth embodiments are denoted by the same symbols. Descriptions of these portions are omitted. Fig. 25, a detection processing device 22 is not shown.

[0182] In the fifth embodiment, an area setting unit 28 of an object recognition device 20 sets a plurality of divided recognition areas AD obtained by dividing a recognition area. Fig. 25 shows a detection area Ad defined in a three-dimensional space area including at least a part of a first side post area, in the case where a shape data acquisition device 21 acquires three-dimensionally controlled shape data. A region setting unit 28 divides the detection area Ad into three divided detection areas ADa, ADb, and ADc. The three divided detection areas ADa, ADb, and ADc are arranged in this order.

[0183] For example, if small objects such as dust are scattered below the shape data acquisition device 21 in a non-divided detection area, the number of pieces of the shape data of the small objects present in the entire detection area exceeds a threshold, and thereby a detection processing unit 29 can detect that a detection target object is present in the detection area. However, the small objects are not objects that can be assumed to be a detection target object. That is, in this case, the detection processing unit 29 may make a false detection.

[0184] By dividing the detection area into different detection areas AD, the number of pieces of small object shape data detected in each divided detection area AD will not easily exceed the threshold. Therefore, the occurrence of false detection can be prevented.

[0185] The recognition processing unit 29 sets a plurality of recognition thresholds for each of the plurality of divided recognition areas ADa, ADb and ADc one to one.

[0186] For example, the plurality of detection thresholds are set so that the farther the distance from the shape data acquisition device 21 is, the smaller the value is. That is, the plurality of detection thresholds are set so that a smaller value is set for a lower divided detection area. As a deeper divided detection area is, the significance of detecting a small object, such as a child's hand, is greater. When an image is acquired as two-dimensionally regulated shape data, the farther it is from the shape data acquisition device 21 and the deeper a divided detection area is, the smaller the image is relative to the entire shape data. That is, the number of pieces of shape data, such as pixels, showing an image of the detection target object decreases relative to the entire image as the distance from the shape data acquisition device 21 increases.

[0187] By setting the detection threshold to be smaller as the divided detection area is smaller, the accuracy of detecting a more significant object can be improved. By setting the detection threshold to be smaller as the divided detection area is smaller, the influence of the number of pieces of shape data, which are smaller depending on the distance from the shape data acquisition device 21, on the detection accuracy can be reduced.

[0188] According to the fifth embodiment described above, the object recognition device 20 defines a plurality of divided detection areas. The object recognition device 20 sets threshold values ​​associated with each of the plurality of divided detection areas. Therefore, the object recognition device 20 can improve the detection accuracy in each divided detection area while simultaneously preventing the occurrence of false detection.

[0189] Next, Fig. 26 describes an example of the hardware constituting the object recognition device 20.

[0190] Fig. 26 is a hardware configuration diagram of the object detection device of the elevator systems of the first to fifth embodiments.

[0191] Each of the functions of the device 20 for detecting objects can be implemented by a processing circuit. For example, the processing circuit is provided with at least one processor 100a and at least one memory 100b. For example, the processing circuit is provided with at least a portion of associated hardware 200.

[0192] When the processing circuit is provided with the at least one processor 100a and the at least one memory 100b, each of the functions of the object recognition device 20 is implemented by software, firmware, or a combination of software and firmware. At least one of the software and firmware is written as a program. At least one of the software and firmware is stored in the at least one memory 100b. The at least one processor 100a implements each of the functions of the object recognition device 20 by reading the program stored in the at least one memory 100b and executing the program. The at least one processor 100a is also referred to as a central processing unit, processing device, computing device, microprocessor, microcomputer, or DSP.The at least one memory 100b is, for example, a non-volatile or volatile semiconductor memory such as a RAM, a ROM, a flash memory, an EPROM or an EEPROM, a magnetic disk, a flexible disk, an optical disk, a compact disc, a minidisc, a DVD or the like.

[0193] If the processing circuit is provided with at least a portion of the associated hardware 200, the processing circuit is implemented, for example, with a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC, an FPGA, or a combination thereof. For example, each of the functions of the object recognition device 20 is implemented by the processing circuit. For example, the functions of the object recognition device 20 are implemented jointly by the processing circuit.

[0194] For each of the functions of the object recognition device 20, one portion thereof may be implemented by the dedicated hardware 200, and the other portion may be implemented by software or firmware. For example, the functions of the recognition processing unit 29 may be implemented by the processing circuit as the dedicated hardware 200, and functions other than the functions of the recognition processing unit 29 may be implemented by the at least one processor 100a, which reads the program stored in the at least one memory 100b and executes the program.

[0195] Thus, the processing circuit implements each of the functions of the object recognition device 20 through the hardware 200, software, firmware, or a combination thereof.

[0196] In the object recognition device 20, the shape data acquisition device 21 and the recognition processing device 22 may be provided at different locations. Each of the functions of the shape data acquisition device 21 and the recognition processing device 22 is also realized by a processing circuit corresponding to the processing circuit that realizes each of the functions of the object recognition device 20, although not shown. Industrial applicability

[0197] As described above, an object detection device according to the present disclosure is applicable to an elevator system. List of reference symbols 1 elevator system, 2 elevator shaft, 3 buildings, 4 engine room, 5 main rope, 6 traction engine, 7 cabins, 7a floor, 7b upper doorpost, 7c, 7d side doorpost, 8 Hall, 9 hall door, 10 hall door frames, 11 Cabin door device, 12 door leaf, 12a single door leaf, 12b, 12c panel, 12d surface, 13 drive engine, 14 control panel, 20 object detection device, 21 Form data acquisition device, 22 recognition processing device, 23 storage unit, 24 receiving unit, 25 range setting unit, 26 first area detection unit, 27 second area detection unit, 28 Detection area setting unit, 29 Detection processing device, 100a processor, 100b memory, 200 hardware QUOTES CONTAINED IN THE DESCRIPTION

[0000] This list of documents submitted by the applicant was generated automatically and is included solely for the convenience of the reader. This list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions. Cited patent literature

[0000] JP 2020-152470 A

[0003]

Claims

[1] An object recognition device comprising: an area setting unit that sets a plurality of areas for shape data of an object located below a shape data detecting device, the shape data detecting device being provided at an upper side of an entrance of a car of an elevator, and the shape data having been detected by the shape data detecting device; a first area detection unit that detects a door area indicating a position of the door from the plurality of areas in the shape data by comparing shape data when a door of the entrance is opened and shape data when the door of the entrance is closed; a second area detection unit that detects one or more second areas from the plurality of areas in the shape data based on a position of the door area detected by the first area detection unit; a detection area setting unit that sets a detection area in at least one of an area including at least a portion of the door area detected by the first area detection unit and an area including at least a portion of the one or more second areas detected by the second area detection unit; and a recognition processing unit that recognizes that a recognition target object exists in the recognition area set by the recognition area setting unit, based on the shape data acquired by the shape data acquisition device. [2] Object recognition device according to claim 1, wherein the shape data is data of a two-dimensional image captured by the shape data capturing device; the area setting unit sets the plurality of areas in the two-dimensional space; and the second area detection unit detects that, as the one or more second areas, two areas adjacent to the door area in an entry / exit direction of the car and present at symmetrical positions in an opening / closing direction of the door, among the plurality of areas set by the area setting unit, are a first side post area and a second side post area indicating positions of paired side posts provided on the car. [3] The object detection device according to claim 2, wherein the detection area setting unit sets at least an area including at least a portion of the first side post area and an area including at least a portion of the second side post area as the detection area. [4] The object detection device according to claim 2 or 3, wherein the detection area setting unit sets an area including at least a portion of the door area detected by the first area detection unit and through which the door passes when the door is opened or closed, as the detection area. [5] The object detection device according to any one of claims 2 to 4, wherein the detection area setting unit sets an area including at least a portion of the door area detected by the first area detection unit and including a position belonging to a hall of the elevator as the detection area. [6] Object recognition device according to one of claims 2 to 5, wherein the detection area setting unit calculates a first area of ​​the first side post area and a second area of ​​the second side post area; if the first area and the second area are the same, respectively sets the area including at least a portion of the first side post area and the area including at least a portion of the second side post area as the detection area; if the first area is smaller than the second area, sets the area including at least the portion of the first side post area as the detection area; and, if the second area is smaller than the first area, sets the area including at least the portion of the second side post area as the detection area. [7] Object recognition device according to one of claims 2 to 5, wherein the detection area setting unit calculating, based on a series of pieces of shape data acquired by the shape data acquiring device during a period from a state in which the entrance door is opened to the closing of the entrance door, displacement amounts between a plurality of pixels constituting the door area of ​​a two-dimensional image included in shape data in the state in which the entrance door is opened and a plurality of pixels constituting the door area of ​​each of a plurality of two-dimensional images included in the series of pieces of shape data; extracts pixels whose displacement amounts exceed a threshold as a plurality of extracted pixels and calculates the center of gravity positions of the plurality of extracted pixels; sets the area including at least the portion of the second side post area as the detection area when it is judged that the first side post area is closer to the center of gravity positions of the plurality of extracted pixels than the second side post area; and sets the area including at least the portion of the first side post area as the detection area when it is judged that the second side post area is closer to the centroid positions of the plurality of extracted pixels than the first side post area. [8] Object recognition device according to claim 1, wherein the shape data is three-dimensional point cloud data, which is a set of measured points in three-dimensional space measured by the shape data acquisition device; the area setting unit detects a plurality of surfaces formed by the three-dimensional point cloud in the three-dimensional space and sets the plurality of areas corresponding to each of the plurality of surfaces; and the second area detection unit recognizes that an area belonging to a surface opposite the shape data acquisition device, among the plurality of areas, is a floor area indicating a position of a floor, as the one or more second areas, and recognizes that two areas orthogonal to the floor area and adjacent to the door area, of the plurality of areas, are a first side post area and a second side post area, indicating the positions of each of the paired side posts provided on the car as the one or more second areas. [9] The object recognition device according to claim 8, wherein the recognition area setting unit sets at least one three-dimensional space area including at least a portion of the first side post area and one three-dimensional space area including at least a portion of the second side post area as the recognition area. [10] The object detection device according to claim 8 or 9, wherein the detection area setting unit sets a three-dimensional space area including at least a portion of the door area detected by the first area detection unit and through which the door passes when the door is opened or closed, as the detection area. [11] Object recognition device according to one of claims 8 to 10, wherein the detection area setting unit sets a three-dimensional space area including at least a portion of the door area detected by the first area detection unit and including a position corresponding to a hall of the elevator as the detection area. [12] The object recognition device according to claim 11, wherein the recognition processing unit calculates a center of gravity position of the object existing within the recognition range based on the three-dimensional point cloud of the shape data, and when it is judged that the object intends to enter the car based on a direction in which the center of gravity position of the object moves, the object is regarded as a recognition target object, and recognizes that the recognition target object exists in the recognition range. [13] Object recognition device according to one of claims 8 to 12, wherein the first area detection unit detects an area parallel to an entry / exit direction of the car from the plurality of areas as a door edge area indicating a boundary position of an inside of the door; and the area setting unit sets a three-dimensional space area including at least a portion of the door edge area as the detection area. [14] Object recognition device according to one of claims 8 to 13, wherein the detection area setting unit calculates a first distance from the shape data acquisition device to the first side post area and a second distance from the shape data acquisition device to the second side post area based on information about the three-dimensional point cloud of the shape data; sets the three-dimensional space area including at least a portion of the first side post area and the three-dimensional space area including at least a portion of the second side post area as the detection area, respectively, when the first distance and the second distance are equal; defines the three-dimensional space area including at least the portion of the first side post area as the detection area when the first distance is greater than the second distance; and sets the three-dimensional space area including at least the portion of the second side post area as the detection area when the second distance is greater than the first distance. [15] The object recognition device according to any one of claims 8 to 14, wherein the recognition processing unit recognizes that the recognition target object exists in the recognition area set by the recognition area setting unit, based on the shape data and measurement intensities at the time of measurement of the measurement points by the shape data acquisition device. [16] An object detection device according to any one of claims 1 to 15, wherein the detection area setting unit causes the shape of the detection area to change depending on the movement of opening or closing the door. [17] Object recognition device according to one of claims 1 to 16, wherein the detection area setting unit sets a plurality of divided detection areas obtained by dividing the detection area; and the recognition processing unit sets a plurality of recognition thresholds corresponding to each of the plurality of divided recognition areas, and recognizes the recognition target object existing in one of the plurality of divided recognition areas based on the shape data acquired by the shape data acquisition device and the plurality of recognition thresholds. [18] An elevator system comprising: a cabin; a door provided at an entrance to the cabin; The object recognition device according to any one of claims 1 to 17, wherein the object recognition device is provided at a top of the entrance of the cabin and includes the shape data acquisition device that acquires shape data of an object located thereunder; and a control panel that stops an opening or closing movement of the door when the object detection device detects that a detection target object is located within a detection area.

Images