JP2014016742A - Fall detection system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a fall detection system capable of reliably detecting a fall of a person by simple processing.SOLUTION: A fall detection system comprises: a distance image sensor 1 that detects a distance value of each pixel in a predetermined detection area; and a fall detection device 3 that detects a fall of a person 4 on the basis of the distance value of each pixel detected by the distance image sensor 1. The fall detection device 3 sets a cuboid 5 based on an outline of the person 4 detected by the distance image sensor 1, and detects a fall of the person 4 on the basis of an aspect ratio of the cuboid 5.

Description

  The present invention relates to a fall detection system, and more particularly to a fall detection system capable of reliably detecting a fall of a person with a simple process using a distance image sensor using a two-dimensional scanner such as a planar actuator. Is.

  In general, escalators, moving walkways, and the like are installed in buildings, station buildings, and the like, and these escalators are usually driven automatically in places where there are no staff. For this reason, when the user falls, there is a problem that it is impossible to respond quickly and it is very dangerous.

  Therefore, even in the past, an escalator monitoring system has been developed in which a camera is installed near the escalator and an image taken by the camera is used to detect a behavior such as a passenger entering or going backward.

  As a technique for detecting such an obstacle, conventionally, for example, at least two cameras are installed in a place where the entire escalator can be monitored, and each image photographed by the cameras is transferred to the image processing apparatus. Then, the image processing device acquires three-dimensional information from each image, analyzes the three-dimensional information, detects the position of the passenger using the escalator, including the depth direction of the imaging range, and operates the escalator. A technique for detecting the behavior of the passenger based on the information is disclosed (for example, see Patent Document 1).

JP 2010-064821 A

  However, since the technique described in Patent Document 1 focuses on the head of a person and detects the fall of the person according to the fluctuation of the height of the head, it is complicated to recognize the head. This necessitates complicated image processing and determination, and thus has a problem that the detection processing becomes complicated. In addition, for example, when a person squats down, there is a possibility that it may be determined that the person falls, and there is a problem that the person's fall cannot be accurately detected. In addition, if a person falls from the height of the head at the start of detection to a height corresponding to a fall for a certain period of time, it is judged that the person has fallen, so the fall analysis must be performed in time series. In addition, the algorithm is complicated.

  Further, in the conventional technique, the person's head is detected by changing from the height at which the person's head starts to fall to the height at which the person's head starts to fall. Therefore, there is a problem in that there is a possibility that the detection processing algorithm may be complicated.

  The present invention has been made in view of the above points, and an object of the present invention is to provide a fall detection system capable of reliably detecting a fall of a person with a simple process.

In order to achieve the above object, according to the present invention, a fall detection system according to the invention of claim 1 includes a distance image sensor that detects a distance value of each pixel in a predetermined detection area, and each of the distance image sensors detected by the distance image sensor. A fall detection device that detects a fall of a person based on a pixel distance value;
The fall detection device sets a rectangular parallelepiped based on the outer shape of the person detected by the distance image sensor, and detects the fall of the person based on the aspect ratio of the rectangular parallelepiped.

  According to a second aspect of the present invention, in the first aspect, the distance image sensor detects a distance value of each pixel by emitting light to the predetermined detection area.

  According to a third aspect of the present invention, in the first or second aspect, the distance image sensor includes a planar actuator and includes a two-dimensional scanner that scans light in a two-dimensional region.

  According to a fourth aspect of the present invention, the fall detection device according to any one of the first to third aspects, wherein the fall detection device focuses on a change in an aspect ratio of only a person's traveling direction from a maximum height of the rectangular parallelepiped. It is characterized by detecting a fall.

  According to a fifth aspect of the present invention, in any one of the first to fourth aspects, the fall detection device sets the width dimension of the rectangular parallelepiped to a constant dimension, and the aspect of the rectangular parallelepiped in which the width dimension is set. It is characterized in that a person's fall is detected based on the ratio.

  The invention according to claim 6 is the fall detection device according to any one of claims 1 to 5, wherein the fall detection device detects a fall of a person by comparing changes in the aspect ratio of the rectangular parallelepiped in time series. It is characterized by.

  According to a seventh aspect of the present invention, in the method according to any one of the first to sixth aspects, the fall detection device has a minimum height having a predetermined interval from the ground among the distance values detected by the distance image sensor. Only the upper distance value is used.

  The invention according to claim 8 further includes a notification device according to any one of claims 1 to 7, further comprising a notification device that notifies the outside when a fall of a person is detected by the fall detection device. Features.

The invention according to claim 9 is the invention according to any one of claims 1 to 8,
The distance image sensor is installed in the vicinity of an escalator, and the fall detection device stops the operation of the escalator when a fall of a person is detected by the fall detection device.

The invention according to claim 10 is any one of claims 1 to 8,
The distance image sensor is installed in the vicinity of a moving sidewalk, and the fall detection device stops the movement of the moving sidewalk when a fall of a person is detected by the fall detection device.

  According to the first aspect of the present invention, the fall detection device sets a rectangular parallelepiped based on the outer shape of the person detected by the distance image sensor, and detects the fall of the person based on the aspect ratio of the rectangular parallelepiped. Therefore, unlike conventional methods, it is not necessary to analyze the person's head and detect the person's fall according to the fluctuation in the height of the head, and it is possible to perform simple processing without complicated image processing and judgment. The fall of a person can be detected reliably. Also, because the person's fall is detected based on the aspect ratio based on the person's external shape, when the person squats down, the aspect ratio does not exceed the threshold value. Can be prevented.

  According to the second aspect of the present invention, since the distance image sensor emits light to a predetermined detection area and detects the distance value of each pixel, based on the distance value detected by the light emission. A person's fall can be detected based on the set aspect ratio of the rectangular parallelepiped.

  According to the third aspect of the invention, the distance image sensor is a planar actuator, and the two-dimensional scanner that scans the light in the two-dimensional region is provided. Therefore, the distance value detected by the light scanning by the planar actuator. The fall of a person can be detected based on the aspect ratio of the rectangular parallelepiped set based on.

  According to the invention of claim 4, the fall detection device detects the fall of the person by focusing on the change in the aspect ratio of only the person's advancing direction from the highest height of the rectangular parallelepiped. Focusing on the aspect ratio toward, it is possible to detect a person falling backward.

  According to the fifth aspect of the present invention, the fall detection device sets the width dimension of the rectangular parallelepiped to a constant dimension, and detects the fall of the person based on the aspect ratio of the rectangular parallelepiped in which the width dimension is set. Therefore, by setting the width dimension of the rectangular parallelepiped, detection of an object other than a person such as a bag held by the person can be eliminated, and the fall of the person can be detected more appropriately.

  According to the invention of claim 6, since the fall detection device compares the change in the aspect ratio of the rectangular parallelepiped in time series and detects the fall of the person, the aspect ratio of the rectangular parallelepiped for each time is compared. Although it is necessary to perform the fall detection using the change in the aspect ratio, more accurate fall detection can be performed.

  According to the invention of claim 7, the fall detection device uses only the distance value above the minimum height having a predetermined interval from the ground among the distance values detected by the distance image sensor. Among the distance values detected by the sensor, unnecessary portions such as the step portion of the escalator can be removed, and an extra distance value can be eliminated to appropriately set the rectangular parallelepiped.

  According to the eighth aspect of the present invention, when a fall of the person is detected by the fall detection device, a notification device for notifying the outside is further provided, so that the fact that the person has fallen is notified to the outside. Can improve safety.

  According to the ninth aspect of the present invention, the distance image sensor is installed in the vicinity of the escalator, and when the fall of the person is detected by the fall detection device, the operation of the escalator is stopped. Can prevent accidents caused by falls and can significantly improve safety.

  According to the invention of claim 10, the distance image sensor is installed in the vicinity of the moving sidewalk, and the movement of the moving sidewalk is stopped when a fall of the person is detected by the fall detection device. It is possible to prevent the accident of a fallen person due to the movement of the sidewalk, and to significantly increase the safety.

It is a schematic structure figure showing an embodiment of a fall detection system concerning the present invention. It is explanatory drawing which shows the setting method of the aspect ratio in embodiment of the fall detection system which concerns on this invention. It is explanatory drawing which shows the state of the rectangular parallelepiped in the state where the person stands in embodiment of the fall detection system which concerns on this invention. It is explanatory drawing which shows the state of the rectangular parallelepiped in the state in which the person is falling down in embodiment of the fall detection system which concerns on this invention. It is explanatory drawing which shows the example at the time of paying attention only to the advancing direction of the person in embodiment of the fall detection system which concerns on this invention. It is an isometric view explanatory drawing which shows the example in the case of setting the width dimension of a rectangular parallelepiped in embodiment of the fall detection system which concerns on this invention. It is plane explanatory drawing which shows the example in the case of setting the width dimension of a rectangular parallelepiped in embodiment of the fall detection system which concerns on this invention. It is a flowchart which shows the operation | movement in embodiment of the fall detection system which concerns on this invention.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 shows a first embodiment of a fall detection system according to the present invention. In this embodiment, a distance image sensor 1 constituting the fall detection system is installed above the entrance of an ascending escalator 2. ing. The distance image sensor 1 is configured to be able to scan the entrance portion of the ascending escalator 2, and the fall detection device 3 is connected to the distance image sensor 1.

  The distance image sensor 1 includes, for example, a two-dimensional scanner (not shown) composed of a planar actuator or the like. The distance image sensor 1 projects laser light onto the two-dimensional scanner while operating the two-dimensional scanner. Each pixel in the range scanned by the two-dimensional scanner is scanned by the two-dimensional scanner by scanning the laser beam in the two-dimensional region and receiving the laser beam projected from the two-dimensional scanner and reflected by the object. The distance value at can be acquired. The distance image sensor 1 is not limited to this. For example, any sensor that can acquire three-dimensional information such as a stereo camera or a sensor that combines an LED and a CMOS can be used. A sensor may be applied.

  In the present embodiment, the fall detection device 3 generates an image based on the distance value of the distance image sensor 1, recognizes an object based on the image, and selects a person 4 from the object. It is configured as follows. The fall detection device 3 is configured to set a rectangular parallelepiped 5 based on the outer shape of the person 4. That is, as shown in FIG. 2, since the height dimension and depth dimension of the person 4 can be detected based on the distance value detected by the distance image sensor 1, it is based on this height dimension and depth dimension. A rectangular parallelepiped 5 is set. And the fall detection apparatus 3 calculates | requires an aspect-ratio based on the height dimension and depth dimension of this rectangular parallelepiped 5. Here, in the present embodiment, the aspect ratio is assumed to be depth dimension (Y) / height dimension (Z).

  The aspect ratio of the rectangular parallelepiped 5 changes when the person 4 falls. Therefore, an aspect ratio at which it can be determined that the person 4 has fallen is set as a threshold value in advance with respect to the aspect ratio when the person 4 is standing, and the fall detection device 3 uses this aspect ratio. When the aspect ratio exceeds the threshold value, it is determined that the person 4 has fallen. For example, as shown in FIG. 3, when the installation height L of the distance image sensor 1 is 3450 mm and the scanning angle θ by the distance image sensor 1 is 60 °, the aspect ratio when the person 4 is standing is <1. However, as shown in FIG. 4, the aspect ratio when the person 4 falls is ≧ 1.7. Therefore, the fall detection device 3 is configured to monitor the aspect ratio of the rectangular parallelepiped 5 and to determine that the person 4 has fallen when the aspect ratio of the rectangular parallelepiped 5 becomes 1.7 or more. The aspect ratio for determining that the person 4 has fallen can be arbitrarily set in consideration of the installation position of the distance image sensor 1, the scanning range, and the like.

  When the fall detection device 3 sets the rectangular parallelepiped 5 of the person 4 based on the distance value of the distance image sensor 1, for example, as shown in FIG. 5, pay attention to the aspect ratio toward the traveling direction of the person 4. Then, the fall of the person 4 may be detected. In this case, the rear of the person 4 is excluded from the aspect ratio calculation. In general, when the person 4 rides on the escalator 2, there is a current situation that the person 4 often falls relatively backward. Therefore, by focusing on the change in the aspect ratio toward the traveling direction of the person 4, the person 4 moves backward. Can be detected.

  Further, when the rectangular parallelepiped 5 based on the outer shape of the person 4 is set by the fall detection device 3, the width of the rectangular parallelepiped 5 is set to a constant dimension, and based on the aspect ratio of the rectangular parallelepiped 5 to which the width dimension is set. The fall of the person 4 may be detected. For example, as shown in FIGS. 6 and 7, the left and right ± 30 cm from the head of the person 4 is set as the width of the rectangular parallelepiped 5, and the aspect ratio is calculated based on the rectangular parallelepiped 5 based on the width. In this way, by setting the width dimension of the rectangular parallelepiped 5, detection of objects other than the person 4, such as a bag held by the person 4, can be eliminated, and the fall of the person 4 can be detected more appropriately. Is. The width dimension of the rectangular parallelepiped 5 can be set to an arbitrary value.

  Further, the fall detection device 3 may compare the change in the aspect ratio of the rectangular parallelepiped 5 in time series and detect the fall of the person 4 based on the aspect ratio that changes with time. In this case, although it is necessary to compare the aspect ratios of the rectangular parallelepiped 5 for each time, the fall detection using the change in the aspect ratio can be performed, so that more accurate fall detection can be performed. .

  Further, the fall detection device 3 uses the distance image sensor 1 to generate an image using only the distance value above the minimum height having a predetermined interval from the ground among the distance values to be detected, and the person 4 based on this image You may make it set the rectangular parallelepiped 5 based on the external shape. In this case, among the distance values detected by the distance image sensor 1, the step portion of the escalator 2 and the like can be removed, and an extra distance value can be eliminated and the rectangular parallelepiped 5 can be set appropriately. Is.

  In the present embodiment, a notification device 6 is connected to the fall detection device 3, and a drive device 7 for the escalator 2 is connected to the fall detection device 3. When the fall detection device 3 detects the fall of the person 4, the fall detection device 3 operates the notification device 6 to notify that the person 4 has fallen around. In this case, the notification means by the notification device 6 may be a voice notification means such as a chime or a light notification means such as a light.

  When the fall detection device 3 detects the fall of the person 4, the drive device 7 is stopped together with the notification by the notification device 6, whereby the escalator 2 is stopped and the person 4 is stopped. It is designed to ensure safety.

  Next, the operation of this embodiment will be described with reference to the flowchart shown in FIG.

  First, when scanning of the laser beam is started by the distance image sensor 1, a distance value is acquired for each pixel in the region scanned with the laser beam (ST1). Then, an image is generated based on the distance value of each pixel, and an object is extracted from this image (ST2). Among the extracted objects, a person 4 is selected from the dimensions of the object (ST3), and a rectangular parallelepiped 5 is set based on the outer shape of the person 4 (ST4).

  Subsequently, the fall detection device 3 calculates the aspect ratio of the rectangular parallelepiped 5 (ST5) and compares it with a preset threshold value (ST6). When the aspect ratio of the rectangular parallelepiped 5 exceeds the threshold value (ST6: YES), it is determined that the person 4 has fallen, the notification device 6 notifies the user (ST7), and the drive device 7 is stopped. (ST9). On the other hand, when the aspect ratio of the rectangular parallelepiped 5 does not exceed the threshold value (ST6: NO), it is determined that the person 4 has not fallen (ST10).

  As described above, in the present embodiment, the rectangular parallelepiped 5 is set based on the outer shape of the person 4, and it is determined that the person 4 has fallen when the aspect ratio of the rectangular parallelepiped 5 exceeds a certain threshold value. Therefore, it is not necessary to analyze the head of the person 4 and detect the fall of the person 4 according to the fluctuation of the height of the head as in the conventional case, and perform complicated image processing and determination. Therefore, it is possible to reliably detect the fall of the person 4 with simple processing. Further, since the fall of the person 4 is detected based on the aspect ratio based on the outer shape of the person 4, when the person 4 crouches, the aspect ratio does not exceed the threshold value. There is no mistaken judgment that you have fallen.

  In the above-described embodiment, the present invention has been described with respect to an example in which the fall detection system according to the present invention is installed on the escalator 2, but the present invention is not limited to this example. The present invention can be applied to any place where the person 4 may fall down.

  Further, the present invention is not limited to the above-described embodiment, and various modifications can be made based on the gist of the present invention.

DESCRIPTION OF SYMBOLS 1 Distance image sensor 2 Escalator 3 Fall detection apparatus 4 Person 5 Rectangular parallelepiped 6 Notification apparatus 7 Drive apparatus

Claims (10)

A distance image sensor that detects a distance value of each pixel in a predetermined detection area, and a fall detection device that detects a person's fall based on the distance value of each pixel detected by the distance image sensor,
The fall detection system is characterized in that the fall detection device sets a rectangular parallelepiped based on the outer shape of the person detected by the distance image sensor, and detects the fall of the person based on the aspect ratio of the rectangular parallelepiped.   The fall detection system according to claim 1, wherein the distance image sensor detects a distance value of each pixel by emitting light to the predetermined detection area.   The fall detection system according to claim 1, wherein the distance image sensor includes a planar actuator and includes a two-dimensional scanner that scans light in a two-dimensional region.   4. The fall detection device according to claim 1, wherein the fall detection device detects a fall of a person by paying attention to an aspect ratio of only a person's traveling direction from a maximum height of the rectangular parallelepiped. 5. Fall detection system.   5. The fall detection device according to claim 1, wherein a width dimension of the rectangular parallelepiped is set to a constant dimension, and a person falls is detected based on an aspect ratio of the rectangular parallelepiped having the width dimension set. The fall detection system according to any one of the above.   The fall detection system according to any one of claims 1 to 5, wherein the fall detection device detects a fall of a person by comparing changes in the aspect ratio of the rectangular parallelepiped in time series. .   The fall detection device uses only a distance value above a minimum height having a predetermined interval from the ground among distance values to be detected by the distance image sensor. The fall detection system according to one item.   The fall detection system according to any one of claims 1 to 7, further comprising a notification device that notifies the outside when a fall of a person is detected by the fall detection device.   The distance image sensor is installed near an escalator, and the fall detection device stops the operation of the escalator when a fall of a person is detected by the fall detection device. The fall detection system according to claim 8.   The distance image sensor is installed in the vicinity of a moving sidewalk, and the fall detection device stops the movement of the moving sidewalk when a fall of a person is detected by the fall detection device. The fall detection system according to any one of claims 1 to 8.

Images