TW201705095A - Apparatus to prevent falling and operation method thereof - Google Patents

Abstract

An apparatus to prevent falling and an operation method thereof are provided. The apparatus includes a depth camera, a control device and a falling buffer device. The depth camera can capture three-dimensional (3D) images of the view field. The control device is coupled to the depth camera to receive the 3D image. The control device can recognize a moving object in the 3D image, and define a platform in the 3D image to obtain a detection area. The control device can determine whether to output a control signal according to the detection area and the moving object. The falling buffer device is coupled to the control device. When the falling buffer device receives the control signal, the falling buffer device can be triggered to move instantly between the moving object and the ground to avoid the moving object falling from the platform.

Description

Drop prevention device and its operation method

The present invention relates to a monitoring device, and more particularly to a method of operating a fall prevention device and a fall prevention device.

Existing sleep monitoring devices have multiple sensors disposed on a mattress. The sleep monitoring device can utilize a sensor to sense whether the user is lying on the bed. Existing sleep monitoring devices must be fitted with a special mattress (a mattress equipped with a sensor) and cannot be applied to a general bed.

On the other hand, existing sleep monitoring devices cannot prevent the user from falling from the bed to the ground. In any event, the user may have been injured when the sleep monitoring device is informed via the sensor that the user has fallen from the bed to the ground.

The present invention provides a method of preventing a drop device and a fall prevention device to prevent a dynamic object from falling from the platform to the ground.

Embodiments of the present invention provide a fall prevention device, including depth photography A depth camera, a control device, and a drop buffer. The depth camera can capture 3D images in the field of view. The control device is coupled to the depth camera to receive the three-dimensional image. The control device can recognize the dynamic object in the 3D image and define the platform in the 3D image to obtain the detection range. The control device can determine whether to generate a control signal according to the detection range and the dynamic object. The drop buffer is coupled to the control device. When the drop buffer receives the control signal, the drop buffer can be triggered to instantly move between the dynamic object and the ground to avoid dropping of the dynamic object from the platform.

In an embodiment of the invention, the control device includes a shooting angle calculation unit, a dynamic object recognition unit, a detection range determination unit, and a determination unit. The shooting angle calculation unit is coupled to the depth camera to receive the three-dimensional image. The shooting angle calculation unit can obtain the shooting direction of the depth camera in the field of view according to the three-dimensional image. The dynamic object recognition unit is coupled to the depth camera and the shooting angle calculation unit. The dynamic object recognition unit can correct the three-dimensional image according to the shooting direction to obtain the corrected three-dimensional image. The dynamic object recognition unit can recognize dynamic objects in the corrected three-dimensional image. The detection range determining unit is coupled to the depth camera and the shooting angle calculation unit. The detection range determining unit can define a platform in the three-dimensional image, and correct the platform according to the shooting direction to obtain the detection range. The determining unit is coupled to the dynamic object recognition unit, the detection range determining unit, and the drop buffer device. The judging unit can compare the relationship between the detection range and the dynamic object to determine whether the control signal is generated to trigger the drop buffer device.

In an embodiment of the invention, the shooting angle calculation unit may search for a plane from the three-dimensional image and estimate the shooting direction of the depth camera according to the plane.

In an embodiment of the invention, the dynamic object recognition unit may extract and establish a background from the three-dimensional image, and extract dynamic objects from the three-dimensional image according to the background.

In an embodiment of the invention, the dynamic object recognition unit can recognize the position of the head, body, upper limb or lower limb of the dynamic object.

In an embodiment of the invention, the dynamic object recognition unit can perform world coordinate conversion on the position of the dynamic object.

In an embodiment of the invention, the detection range determining unit may determine the level of the detection range according to a platform defined by the user from the three-dimensional image.

In an embodiment of the invention, the detection range determining unit may determine the horizontal heights of the plurality of planes in the three-dimensional image, and define one of the planes as the platform according to the horizontal heights, and according to the shooting direction. Fix the platform to get the detection range.

In an embodiment of the invention, the detection range determining unit may perform world coordinate conversion on the position of the detection range.

In an embodiment of the invention, the above-mentioned fall prevention device further includes a fall warning device. The fall warning device is coupled to the control device. When the fall warning device is triggered, the fall warning device immediately prevents the dynamic object from falling from the platform. Among them, when the position of the body of the dynamic object is within the detection range, and the head of the dynamic object exceeds When the detection range and the height of the head of the dynamic object are lower than the lying range, the determining unit generates the control signal to trigger the drop warning device.

Embodiments of the present invention provide an operation method for preventing a drop device, including Included: receiving a three-dimensional image from a depth camera to capture a field of view; identifying a dynamic object in the three-dimensional image through a control device, and defining a platform in the three-dimensional image to obtain a detection range; determining the control device according to the detection range and the dynamic object Whether a control signal is generated to trigger the drop buffer device; and when the drop buffer device receives the control signal, the drop buffer device is instantly moved between the dynamic object and the ground to prevent the dynamic object from falling from the platform.

In an embodiment of the present invention, the foregoing operating method further includes: controlling The shooting angle calculation unit of the device obtains the shooting direction of the depth camera in the field of view according to the three-dimensional image; the dynamic object recognition unit of the control device corrects the three-dimensional image according to the shooting direction to obtain the corrected three-dimensional image, and recognizes the dynamic in the corrected three-dimensional image. The object is defined by the detection range determining unit of the control device, and the detection range is obtained by modifying the platform according to the shooting direction; and the determining unit of the control device compares the detection range with the dynamic object, Decide whether to trigger the drop buffer.

In an embodiment of the invention, the obtaining the depth camera is as described above The step of the shooting direction in the field of view includes: finding a plane from the three-dimensional image; and estimating the shooting direction of the depth camera according to the plane.

In an embodiment of the invention, the modified three-dimensional image is obtained to obtain a The steps of correcting the three-dimensional image include: converting the three-dimensional image into a world coordinate to obtain Corrected 3D images.

In an embodiment of the invention, the step of identifying the dynamic object in the corrected three-dimensional image comprises: identifying the position of the head, body, upper limb or lower limb of the dynamic object.

In an embodiment of the invention, the step of obtaining the detection range includes: determining, according to a platform defined by the user from the three-dimensional image, a level of the detection range of the platform; and determining a position of the detection range Perform world coordinate conversion.

In an embodiment of the invention, the step of obtaining the detection range The method includes: determining a horizontal level of a plurality of planes in the three-dimensional image; defining one of the planes as the platform according to the horizontal heights; and correcting the platform according to the shooting direction to obtain a detection range.

In an embodiment of the invention, the step of obtaining the detection range by modifying the platform comprises: performing global coordinate conversion on the position of the detection range.

In an embodiment of the invention, the determining unit compares the detection range with the dynamic object in a world coordinate relationship to determine whether the control signal is generated to trigger the drop buffer device.

In an embodiment of the invention, when the position of the body of the dynamic object is within the detection range, and the position of the center of gravity of the body of the dynamic object exceeds the detection range, and the height of the head and the body of the dynamic object is lower than the lying range. When the control unit generates the control signal to trigger the drop buffer device.

In an embodiment of the invention, the drop buffer device comprises a self-propelled cushion or an airbag.

In an embodiment of the invention, the foregoing operation method further comprises: when preventing When the fall warning device of the fall arrest device is triggered, the fall warning device immediately prevents the dynamic object from falling from the platform.

In an embodiment of the invention, when the position of the body of the dynamic object is in the detective Within the measurement range, and when the head of the dynamic object exceeds the detection range, and the height of the head of the dynamic object is lower than the lying range, the fall warning device is triggered by the above-mentioned judging unit.

In an embodiment of the invention, the drop warning device comprises a self-lifting Bed edge guardrail.

In an embodiment of the invention, the drop buffer device further includes a police A reporter that emits an alarm to indicate that a dynamic object is about to fall from the platform.

Based on the above, the drop preventing device and the operator thereof are described in the embodiment of the present invention. The method uses a depth camera to monitor the relationship between dynamic objects and the platform, so it can be applied to a general environment without the need to cooperate with a special platform. Furthermore, the fall prevention device and the operation method thereof according to the embodiments of the present invention use the drop buffer device to instantly move between the dynamic object and the ground, thereby preventing the dynamic object from falling from the platform to the ground.

In order to make the above features and advantages of the present invention more apparent, the following is a special The embodiments are described in detail below in conjunction with the drawings.

100‧‧‧Drop prevention device

110‧‧‧Deep camera

111‧‧ Sight

120‧‧‧Control device

121‧‧‧Dynamic object recognition unit

122‧‧‧ shooting angle calculation unit

123‧‧‧Detection range decision unit

124‧‧‧judging unit

130‧‧‧Drop buffer

140‧‧‧Drop warning device

210‧‧‧ bed

220‧‧‧ people

230‧‧‧ feet

240‧‧‧ pillar

241‧‧‧Upper end

S310~S340, S610~S650, S702~S744‧‧‧ steps

1 is a block diagram illustrating a fall prevention device in accordance with an embodiment of the present invention; schematic diagram.

FIG. 2 is a schematic diagram showing an application scenario of the fall prevention device shown in FIG. 1 according to an embodiment of the invention.

3 is a flow chart showing a method of preventing a drop device in accordance with an embodiment of the present invention.

4A and 4B are schematic diagrams showing an application scenario of a drop buffer device according to an embodiment of the invention.

5A and 5B are schematic diagrams showing an application scenario of a drop buffer device according to another embodiment of the present invention.

FIG. 6 is a flow chart showing a method of preventing a drop device according to another embodiment of the present invention.

FIG. 7 is a flow chart showing a method of preventing a drop device according to still another embodiment of the present invention.

"Coupling" used in the full text of this prospectus (including the scope of patent application) The term "接接" can refer to any direct or indirect means of attachment. For example, if the first device is described as being coupled to the second device, it should be construed that the first device can be directly connected to the second device, or the first device can be connected through other devices or some kind of connection means. Connected to the second device indirectly. In addition, wherever possible, the elements and/ Components/components that use the same reference numbers or use the same terminology in different embodiments The steps can refer to the relevant instructions.

1 is a block diagram showing a circuit for preventing a drop device 100 in accordance with an embodiment of the present invention. The fall prevention device 100 includes a depth camera 110, a control device 120, and a drop buffer device 130. In some implementations, the depth camera 110 can be placed on a wall or on a ceiling. In other implementations, the depth camera 110 can be placed on a headboard (or on a bed rail) or on a footboard (or on a tail rail). In still other implementations, the depth camera 110 can be disposed on the control device 120. The depth camera 110 can capture a three-dimensional image within the field of view. There may be beds and people in the field of view.

For example, but not limited to, FIG. 2 is a schematic diagram illustrating an application scenario of the fall prevention device 100 of FIG. 1 in accordance with an embodiment of the present invention. In the embodiment shown in FIG. 2, control device 120 is disposed within foot 230. A pillar 240 is provided on the foot 230, and a depth camera 110 is placed on the upper end portion 241 of the pillar 240. The depth camera 110 and the control device 120 can be moved to any location according to application requirements. The depth camera 110 can capture a three-dimensional image within the field of view 111. There may be a platform (e.g., bed 210) and a dynamic object (e.g., human 220 or other animal) within field of view 111.

3 is a flow chart showing a method of preventing a drop device in accordance with an embodiment of the present invention. Referring to FIG. 1 and FIG. 3, in step S310, the depth camera 110 can continuously capture the three-dimensional image in the field of view 111. That is, the depth camera 110 can continuously provide the three-dimensional image to the control device 120. This embodiment is not Embodiments of the depth camera 110 are limited. For example, in some embodiments, depth camera 110 may include multiple cameras. The depth camera 110 can determine the distance from the depth camera 110 to the object based on the difference in viewing angles of the plurality of imaging devices. The depth camera 110 can use a conventional stereo image method or other algorithms to obtain a three-dimensional image.

In other embodiments, the depth camera 110 may employ an active depth map sensor. For example, the depth camera 110 may use a light-section method (ie, the distance from the depth camera 110 to an object based on the principle of triangulation), or a time-of-flight method. The method is to measure the time difference between the projection light projected onto the object and the reflected light reflected from the object. In other embodiments, the depth camera 110 may project a plurality of light patterns having different spatial phases, and determine a three-dimensional shape of the surface of the object based on a positional relationship of the light patterns formed on the surface of the object.

The control device 120 is coupled to the depth camera 110 to receive the three-dimensional image. In step S320, the control device 120 can recognize the dynamic object in the three-dimensional image. For example, when the person 220 shown in FIG. 2 is turned over on the bed 210, the control device 120 can recognize the person 220 (dynamic object) that is moving in the three-dimensional image.

In step S330, the control device 120 may define a platform in the three-dimensional image to obtain a detection range. In some embodiments, the control device 120 can provide a user interface, and the user can arbitrarily define a certain platform in the three-dimensional image (eg, the bed 210 shown in FIG. 2). The control device 120 can determine/obtain the detection range according to the platform defined by the user. According to the application requirements, the detection range can be equivalent to three-dimensional shadow The full extent of the platform in the image (such as the full horizontal bed of the bed 210 shown in Figure 2), or the detection range may be slightly smaller than the platform. For example, the control device 120 can retract the edge of the platform in the three-dimensional image a safe distance inward as the detection range. The safety distance may be determined according to application requirements, for example, the safety distance is set to 3 cm, 10 cm or other distance.

In other embodiments, control device 120 can identify all levels in the three-dimensional image and define the lowest level as the ground. Based on ergonomics, the bed height is often regulated in a bed height range. Thus, control device 120 can determine the distance from each horizontal plane to the ground (i.e., the height of the horizontal plane) in the three-dimensional image, and define the level that corresponds to the height of the bed as the platform (e.g., bed 210 shown in FIG. 2). Further, the control device 120 can determine/obtain the detection range according to the automatically defined platform. The determination of the detection range can be referred to the description of the previous paragraph and so on, so it will not be described again.

The drop buffer device 130 is coupled to the control device 120. In step S340, the control device 120 may determine whether to generate a control signal to trigger the drop buffer device 130 according to the relationship between the detection range and the dynamic object. When a dynamic object (such as the person 220 shown in FIG. 2) is about to fall from the detection range (eg, the bed 210 shown in FIG. 2) to the ground, the control device 120 can generate a control signal to the drop buffer 130. When the drop buffer 130 receives the control signal, the drop buffer 130 can be triggered to instantly move between the dynamic object and the ground to avoid dropping of the dynamic object from the platform.

This embodiment does not limit the implementation of the drop buffer device 130. Example In other embodiments, the drop buffer 130 may include a self-propelled cushion. 4A and 4B are schematic diagrams showing an application scenario of the drop buffer device 130 according to an embodiment of the invention. The drop buffer device 130 (self-propelled cushion) is provided with a caster, a traveling power mechanism, and a cushion. The material of the cushion may include a spring mattress, a foam pad, a latex pad or other soft/elastic materials. At normal times, the drop buffer 130 remains in the stowed position, such as below the bed 210, as shown in Figure 4A.

Referring to FIG. 4B, when the control device 120 detects that the person 220 (dynamic object) protrudes from the edge of the bed 210 by the depth camera 110 and is about to fall from the bed 210 (detection range) to the ground, the control device 120 can immediately The drop buffer device 130 is triggered. When the drop buffer device 130 is triggered, the drop buffer device 130 can be quickly moved from the stowed position shown in FIG. 4A to the receiving position shown in FIG. 4B. Therefore, the drop buffer device 130 can instantly block the person 220 who has fallen from the bed 210. The drop buffer 130 can prevent the person 220 from falling from the bed 210 to the ground. In some embodiments, but not limited thereto, the drop buffer 130 may further include an alarm to alert that the person 220 (dynamic article) has fallen from the bed 210 (platform).

In some application examples, the control device 120 can further report the monitoring content to the remote device by using a wireless communication method (or a wired communication network). For example, but not limited to, the control device 120 can communicate with the remote device by using a Local Area Network (LAN), a nurse call system communication network, or other communication interface. In some application scenarios, the fall prevention device 100 may be applied to a hospital ward, an elderly facility or a care facility, and the like. For example in a hospital In the ward, when the control device 120 detects that the person 220 (dynamic object) is about to fall from the bed 210 (detection range) to the ground through the depth camera 110, the control device 120 can also be set in addition to the trigger buffer device 130. The remote device of the hospital's nurse center issued an alarm to inform the medical staff of the drop event.

In other embodiments, but not limited to, the cushion of the drop buffer device 130 can be configured with one or more sensors. When the triggered drop buffer device 130 detects that the cushion has been received by the person 220 (dynamic object), the drop buffer device 130 will remain in the receiving position shown in FIG. 4B, and the control device 120 and/or the drop buffer device 130 may An alert is issued to notify the caregiver. When the triggered drop buffer device 130 does not detect the person 220 (dynamic object), it indicates that the person 220 may fall on the ground (or the person 220 climbs back to the bed 210 from the drop buffer device 130), so the drop buffer device 130 is from FIG. 4B. The illustrated receiving position automatically returns to the stowed position shown in Figure 4A and an alert is issued to notify the caregiver.

In other embodiments, the drop buffer 130 may include an airbag (like an anti-collision airbag in a vehicle cockpit). 5A and 5B are schematic diagrams showing an application scenario of the drop buffer device 130 according to another embodiment of the present invention. The drop buffer device 130 (airbag) has a capsule and an automatic inflation mechanism. At normal times, the drop buffer 130 remains in an uninflated state and is received below the edge of the bed 210, as shown in Figure 5A.

Referring to FIG. 5B, when the control device 120 detects through the depth camera 110 that the person 220 (dynamic object) protrudes from the edge of the bed 210, and is about to be taken from the bed 210 (detection) When the measurement range is dropped to the ground, the control device 120 can immediately trigger the drop buffer device 130. When the drop buffer 130 is triggered, the drop buffer 130 can be quickly inflated to become an air bag (similar to an air mattress), as shown in Figure 5B. Therefore, the drop buffer device 130 can instantly block the person 220 who has fallen from the bed 210. The drop buffer 130 can prevent the person 220 from falling from the bed 210 to the ground. In some embodiments, but not limited thereto, the drop buffer 130 may further include an alarm to alert that the person 220 (dynamic article) has fallen from the bed 210 (platform).

In some embodiments, but not limited to, the fall prevention device 100 illustrated in FIG. 1 may further include a fall warning device 140. The drop buffer device 130 and/or the fall warning device 140 may be selectively disposed to prevent the drop device 100. The fall warning device 140 is coupled to the control device 120. When the control device 120 detects that a dynamic object (for example, the person 220 shown in FIG. 2) approaches the edge of the detection range (for example, the edge of the bed 210 shown in FIG. 2) by the depth camera 110, it indicates that the dynamic object falls from the detection range to The probability of the ground is greatly increased, that is, the dynamic object is very likely to fall from the detection range to the ground. At this time, the control device 120 can immediately trigger the fall warning device 140.

When the fall warning device 140 of the fall prevention device 100 is triggered, the fall warning device 140 can immediately prevent a dynamic object (such as the person 220 shown in FIG. 2) from falling from the platform (eg, the bed 210 shown in FIG. 2). In some embodiments, the fall warning device 140 may include a self-elevating bed edge barrier. The fall warning device 140 (self-elevating bed edge guardrail) can be placed at the edge of the platform (e.g., the bed edge of the bed 210 shown in Figure 2). The fall warning device 140 has a guardrail (or baffle) and a lifting power mechanism. In normal times, the fall warning device 140 can be held in the storage position, for example as shown in FIG. Below the bed 210.

When the control device 120 detects that the person 220 (dynamic object) is approaching the edge of the bed 210 (detection range) by the depth camera 110 (the person 220 may have to fall from the bed 210), the control device 120 can immediately trigger the drop warning device 140. When the fall warning device 140 is triggered, the fall warning device 140 can quickly rise from the stowed position to the blocked position to become the bed edge guard of the bed 210 to protect the person 220. Thus, the fall warning device 140 can immediately prevent/prevent the fall of a dynamic object (such as the person 220 shown in FIG. 2) from the platform (eg, the bed 210 shown in FIG. 2).

In some alternative embodiments, when the fall warning device 140 is triggered, the control device 120 and/or the fall warning device 140 can issue an alert to notify the caregiver. For example, in addition to triggering the fall warning device 140, the control device 120 can also alert the remote device.

In some embodiments, but not limited thereto, the control device 120 shown in FIG. 1 includes a dynamic object recognition unit 121, a shooting angle calculation unit 122, a detection range determination unit 123, and a determination unit 124. The dynamic object recognition unit 121, the shooting angle calculation unit 122, and the detection range determining unit 123 are coupled to the depth camera 110 to receive the three-dimensional image.

FIG. 6 is a flow chart showing a method of preventing a drop device according to another embodiment of the present invention. Referring to FIG. 1 and FIG. 6, in step S610, the depth camera 110 can continuously capture the three-dimensional image in the field of view 111, and continuously provide the three-dimensional image to the dynamic object recognition unit 121, the shooting angle calculation unit 122, and the detection range determining unit 123. . Step S610 shown in FIG. 6 can refer to step S310 shown in FIG. Related instructions and so on. In step S620, the shooting angle calculation unit 122 can obtain the shooting direction (photographing angle) of the depth camera 110 in the field of view 111 according to the three-dimensional image.

The dynamic object recognition unit 121 is coupled to the shooting angle calculation unit 122. In step S630, the dynamic object recognition unit 121 may correct the three-dimensional image according to the shooting direction of the depth camera 110 to obtain the corrected three-dimensional image, and identify the dynamic object in the corrected three-dimensional image (for example, the person 220 shown in FIG. 2). Step S630 shown in FIG. 6 can be analogized with reference to the related description of step S320 shown in FIG.

The detection range determining unit 123 is coupled to the shooting angle calculation unit 122. In step S640, the detection range determining unit 123 can define a platform in the three-dimensional image, and correct the platform according to the shooting direction of the depth camera 110 to obtain the detection range. Step S640 shown in FIG. 6 can be analogized with reference to the related description of step S330 shown in FIG.

The determining unit 124 is coupled to the dynamic object identifying unit 121, the detecting range determining unit 123, and the drop buffering device 130. In step S650, the determining unit 124 may compare the relationship between the detection range and the dynamic object to determine whether a control signal is generated to trigger the drop buffer device 130. Step S650 shown in FIG. 6 can be analogized with reference to the related description of step S340 shown in FIG.

FIG. 7 is a flow chart showing a method of preventing a drop device according to still another embodiment of the present invention. Referring to FIG. 1 and FIG. 7, in step S702, the depth camera 110 can continuously capture the three-dimensional image in the field of view 111, and continuously provide the three-dimensional image to the dynamic object recognition unit 121, the shooting angle calculation unit 122, and the detection. Range decision unit 123. Step S702 shown in FIG. 7 can be analogized with reference to the related description of step S310 shown in FIG.

In step S704, the shooting angle calculation unit 122 may search for one or more planes (for example, a floor, a wall, a ceiling, and/or other planes in the three-dimensional image) from the three-dimensional image. In step S706, the shooting angle calculation unit 122 may estimate the shooting direction (photographing angle) of the depth camera 110 in accordance with the plane of step S704. For example, a conventional bed (planar) is square in plan view; when the bed is squinted at a certain angle of view, the bed may exhibit a trapezoidal or other deformation. The degree of deformation of the plane is related to the photographing direction (photographing angle) of the depth camera 110. Therefore, the photographing angle calculation unit 122 can recognize/infer the photographing direction (the photographing angle) of the depth camera 110 from the degree of deformation of the plane in the image.

In step S708, the dynamic object recognition unit 121 can extract and establish a background from the three-dimensional image. For example, the dynamic object recognition unit 121 can compare a plurality of images taken at different times, and extract the unchanging portions (the same portion) between the images as a background.

In step S710, the dynamic object recognition unit 121 can determine whether there is a dynamic object (for example, the person 220 shown in FIG. 2) in the three-dimensional image according to the background of step S708. For example, the dynamic object recognition unit 121 may compare the background of step S708 with the three-dimensional image provided by the depth camera 110, and extract the object different from the background of step S708 as a dynamic object. If it is determined in step S710 that there is no dynamic object in the three-dimensional image, the process returns to step S708. If it is determined in step S710 that there is a dynamic object in the three-dimensional image, step S712 is performed.

In step S712, the dynamic object recognition unit 121 may extract a dynamic object (for example, the person 220 shown in FIG. 2) from the three-dimensional image according to the background of step S708. According to the shooting direction (photographing angle) of the depth camera 110 obtained in step S706, the dynamic object recognition unit 121 can also correct the three-dimensional image provided by the depth camera 110 in step S712 to correct the three-dimensional image because the shooting direction of the depth camera 110 (photographing) The deformation caused by the angle). On the other hand, the dynamic object recognition unit 121 may perform a coordinate system conversion of the three-dimensional image provided by the depth camera 110, for example, from the camera coordinates to the world coordinates, in step S712 to obtain a corrected three-dimensional image. The coordinate system conversion can be converted by coordinates of the conventional image processing method, and therefore will not be described again. Therefore, the dynamic object recognition unit 121 can perform world coordinate conversion of the position of the dynamic object (for example, the person 220 shown in FIG. 2).

In step S714, the dynamic object recognition unit 121 may determine whether the dynamic object captured in step S712 is a human. For example, the dynamic object recognition unit 121 can recognize whether the dynamic object has a head, a hair portion, a body portion, and/or a limb portion, thereby determining whether the dynamic object captured in step S712 is a human. For example, the dynamic object recognition unit 121 can recognize whether the dynamic object has an eye, a nose, a mouth, and/or an ear, thereby determining whether the dynamic object captured in step S712 is a human. If it is determined in step S714 that the dynamic object is not a human, then the process returns to step S710. If it is determined in step S714 that the dynamic object is a person, step S716 is performed.

In step S716, the dynamic object recognition unit 121 can recognize the positions of the head, body, upper limbs, and/or lower limbs of the dynamic object (such as the person 220 shown in FIG. 2). In step S716, the dynamic object recognition unit 121 can mark and continuously track the location. Head, body, upper limbs and/or lower limbs. For example, but not limited to, the dynamic object recognition unit 121 can use the conventional Open Natural Interaction (OpenNI) technology to identify the position of the head, body, upper limbs, and/or lower limbs of the dynamic object. And labeling and continuously tracking the head, body, upper limbs and/or lower limbs.

In step S720, the detection range determining unit 123 can determine whether the user defines the platform from the three-dimensional image. For example, the control device 120 can provide a user interface, and the user can arbitrarily define a certain platform in the three-dimensional image (for example, the bed 210 shown in FIG. 2). If it is determined in step S720 that the user has defined the platform from the three-dimensional image, the process proceeds to step S722. If it is determined in step S720 that the user does not define the platform from the three-dimensional image, step S724 is performed.

In step S722, the detection range determining unit 123 can determine the level of the detection range according to the platform defined by the user from the three-dimensional image. For example, but not limited to, the detection range determining unit 123 can retract the user from the edge of the platform defined in the three-dimensional image (for example, the bed 210 shown in FIG. 2) by a safe distance as the Detector. Measuring range. The safety distance may be determined according to application requirements, for example, the safety distance is set to 3 cm, 10 cm or other distance. Next, the detection range determining unit 123 can determine the level of the detection range.

In step S724, the detection range determining unit 123 can determine the level of the plurality of planes in the three-dimensional image. In step S726, the detection range determining unit 123 can define one of the planes from the planes according to the horizontal heights. A defined platform (such as bed 210 shown in Figure 2). For example, but not limited to, the detection range determining unit 123 can recognize all planes in the three-dimensional image (for example, floor, bed, table, or other horizontal plane), calculate the height of the planes, and calculate the lowest plane. Defined as ground. Based on ergonomics, the bed height is often regulated in a bed height range. Therefore, the detection range determining unit 123 can determine the distance from each plane of the three-dimensional image to the ground (ie, the horizontal height), and automatically adopt the plane whose horizontal height conforms to the bed height range as the defined platform (for example, the bed shown in FIG. 2) 210). Further, the control device 120 can determine/obtain the detection range according to the automatically defined platform. The determination of the detection range can be referred to the description of the previous paragraph and so on, so it will not be described again. In other embodiments, the detection range determining unit 123 may further correct the defined platform according to the shooting direction (the shooting angle) of the depth camera 110 obtained in step S706 to obtain the detection range.

In step S728, the detection range determining unit 123 can perform the world coordinate conversion on the position of the detection range. For example, the detection range determining unit 123 can convert the position of the detection range from the camera coordinates to the world coordinates. The coordinate system conversion can be converted by coordinates of the conventional image processing method, and therefore will not be described again.

The determining unit 124 may compare the detection range with the dynamic object in the world coordinate (step S730) to determine whether to generate a control signal to trigger the drop buffer device 130 and/or the fall warning device 140. In step S732, the determining unit 124 can determine whether the position of the body of the dynamic object (for example, the person 220 shown in FIG. 2) is within the detection range (for example, the bed 210 shown in FIG. 2). If the step S732 determines that the position of the body of the dynamic object is not in the detection range (for example, the person 220 is not lying in the bed) Returning to step S730. If it is determined in step S732 that the position of the body of the dynamic object is within the detection range, then steps S734 and S740 are performed.

In step S734, the determining unit 124 can determine whether the head of the dynamic object (for example, the person 220 shown in FIG. 2) exceeds the detection range. If it is determined in step S734 that the head of the dynamic object does not exceed the detection range (for example, the bed 210 shown in FIG. 2), the process returns to step S730. If it is determined in step S734 that the head of the dynamic object is out of the detection range, then step S736 is performed.

In step S736, the determining unit 124 can determine whether the height of the head of the dynamic object (for example, the person 220 shown in FIG. 2) is lower than the lying range. The size of the heads based on different people is not much different from each other. When the head is lying on the bed, the height difference between the head and the bed surface (the height of the head) is often regulated in a lying range. Therefore, the judging unit 124 can judge whether the person is lying on the bed according to the relationship between the height of the head (relative to the bed surface) and the lying range. If it is determined in step S736 that the height of the head of the dynamic object is not lower than the lying range, then the flow returns to step S730. If it is determined in step S736 that the height of the head of the dynamic object is lower than the lying range, step S738 is performed.

In step S738, the determining unit 124 can trigger the fall warning device 140. In some embodiments, the fall warning device 140 includes a self-elevating bed edge barrier. When the fall warning device 140 is triggered, the jack-up bed guard rail can instantly prevent a dynamic object (such as the person 220 shown in FIG. 2) from falling from the platform (eg, the bed 210 shown in FIG. 2). In other embodiments, the fall warning device 140 includes an alarm. When the fall warning device 140 is triggered, the alarm can sound an alarm to indicate that the dynamic object is about to fall from the platform.

In step S740, the determining unit 124 can determine whether the position of the center of gravity of the body of the dynamic object (for example, the person 220 shown in FIG. 2) exceeds the detection range (for example, the bed 210 shown in FIG. 2). If it is determined in step S740 that the center of gravity of the body of the dynamic object does not exceed the detection range, then the process returns to step S730. If it is determined in step S740 that the position of the center of gravity of the body of the dynamic object exceeds the detection range, step S742 is performed.

In step S742, the determining unit 124 can determine whether the height and position of the head and the body of the dynamic object (for example, the person 220 shown in FIG. 2) are abnormal. For example, the judging unit 124 can judge whether the height of the head and the body of the dynamic object is lower than the lying range. When the head is lying on the bed, the height difference (or height) between the head (body) and the bed surface of the different person is not much different from each other, and it can be standardized in a lying range. Therefore, the judging unit 124 can judge whether the person is lying on the bed according to the relationship between the height of the head and the body (relative to the bed surface) and the lying range. If it is determined in step S742 that the height of the head and the body of the dynamic object is not lower than the lying range, then step S738 is performed. If it is determined in step S742 that the height of the head and the body of the dynamic object is lower than the lying range, step S744 is performed.

In step S744, the determining unit 124 may generate a control signal to trigger the drop buffer device 130. In some embodiments, the drop buffer device 130 includes a self-propelled cushion. When the drop buffer 130 is triggered, the self-propelled cushion can be moved instantly between the dynamic object (such as the person 220 shown in Figure 2) and the ground. In other embodiments, the drop buffer 130 includes an airbag (like an anti-collision airbag in a vehicle cockpit). When the drop buffer 130 is triggered, the airbag can be quickly inflated to become an air bag to block dynamic objects that fall from the platform. In still other embodiments, the drop buffer device 130 includes an alarm. When the drop buffer 130 is triggered, the alarm can sound an alarm to indicate that the dynamic object has fallen from the platform.

In summary, the fall prevention device 100 and the method of operating the same according to embodiments of the present invention utilize the depth camera 110 to monitor mutual interaction between a dynamic object (such as the person 220 shown in FIG. 2) and a platform (such as the bed 210 shown in FIG. 2). The relationship can therefore be applied to the general environment without the need to match a special mattress (a mattress equipped with a sensor). Moreover, the fall prevention device 100 and the method of operating the same according to the embodiments of the present invention use the drop buffer device 130 to instantly move between the dynamic object and the ground, thereby preventing the dynamic object from falling from the platform to the ground.

Although the present invention has been disclosed in the above embodiments, it is not intended to limit the present invention, and any one of ordinary skill in the art can make some changes and refinements without departing from the spirit and scope of the present invention. The scope of the invention is defined by the scope of the appended claims.

110‧‧‧Deep camera

111‧‧ Sight

120‧‧‧Control device

130‧‧‧Drop buffer

210‧‧‧ bed

220‧‧‧ people

230‧‧‧ feet

240‧‧‧ pillar

241‧‧‧Upper end

Claims (27)

A fall prevention device includes: a depth camera for capturing a three-dimensional image in a field of view; a control device coupled to the depth camera to receive the three-dimensional image, identifying a dynamic object in the three-dimensional image, and defining the Obtaining a detection range from a platform in the 3D image, and determining whether to generate a control signal according to the detection range and the dynamic object; and a drop buffer device coupled to the control device, wherein the drop buffer The device receives the control signal, triggering the drop buffer device to instantly move between the dynamic object and a ground to prevent the dynamic object from falling from the platform. The anti-drop device of claim 1, wherein the control device comprises: a shooting angle calculation unit coupled to the depth camera to receive the three-dimensional image, and the depth camera is obtained in the field of view according to the three-dimensional image a dynamic object recognition unit coupled to the depth camera and the shooting angle calculation unit, correcting the three-dimensional image according to the shooting direction to obtain a corrected three-dimensional image, and identifying the dynamic in the corrected three-dimensional image a detection range determining unit coupled to the depth camera and the shooting angle calculation unit, defining the platform in the three-dimensional image, and modifying the platform according to the shooting direction to obtain the detection range; and a determining unit And coupled to the dynamic object recognition unit, the detection range determining unit, and the drop buffer device, compared to the detection range and the dynamic object Relationship to determine whether the control signal is generated to trigger the drop buffer. The fall prevention device of claim 2, wherein the shooting angle calculation unit searches for a plane from the three-dimensional image, and estimates the shooting direction of the depth camera according to the plane. The fall prevention device of claim 2, wherein the dynamic object recognition unit extracts and establishes a background from the three-dimensional image, and extracts the dynamic object from the three-dimensional image according to the background. The fall prevention device of claim 4, wherein the dynamic object recognition unit recognizes a position of a head, a body, an upper limb or a lower limb of the dynamic object. The fall prevention device of claim 4, wherein the dynamic object recognition unit converts the position of the dynamic object to a world coordinate. The fall prevention device of claim 2, wherein the detection range determining unit determines the level of the detection range according to the platform defined by the user from the three-dimensional image. The anti-drop device of claim 2, wherein the detection range determining unit is configured to determine a level of a plurality of planes in the three-dimensional image, and define from the planes according to the level of the planes One of the planes serves as the platform, and the detection range is obtained by modifying the platform according to the shooting direction. The anti-drop device of claim 8, wherein the detection range determining unit is configured to perform world coordinate conversion on the position of the detection range. The fall prevention device of claim 2, wherein the determining unit compares the detection range with the dynamic object at a world coordinate, The drop buffer is triggered to determine whether the control signal is generated. The fall prevention device of claim 10, wherein when the position of the body of the dynamic object is within the detection range, and the position of the center of gravity of the body of the dynamic object exceeds the detection range, and the dynamic object When the height of the head and the body is lower than a lying range, the determining unit generates the control signal to trigger the drop buffer device. The fall prevention device of claim 1, wherein the drop buffer device comprises a self-propelled cushion or an airbag. The fall prevention device of claim 1, further comprising: a fall warning device coupled to the control device, wherein the fall warning device immediately blocks the dynamic object from being triggered when the drop warning device is triggered a fall of the platform; wherein when the position of the body of the dynamic object is within the detection range, and the head of the dynamic object exceeds the detection range, and the height of the head of the dynamic object is lower than a lying range, The determining unit generates the control signal to trigger the drop warning device. The fall prevention device of claim 13, wherein the fall warning device comprises a self-lifting bed edge guard; wherein the fall warning device comprises an alarm for issuing an alarm to indicate that the dynamic object is about to be The platform fell. An operation method for preventing a fall device includes: receiving a three-dimensional image in a field of view from a depth camera; identifying a dynamic object in the three-dimensional image through a control device, and defining a platform in the three-dimensional image to obtain a Detection range; Determining whether the control device generates a control signal to trigger a drop buffer device according to the detection range and the dynamic object; and when the drop buffer device receives the control signal, immediately moving the drop buffer device to the dynamic Between the object and a ground to avoid the dynamic object falling from the platform. The method for operating the fall prevention device of claim 15 further includes: obtaining, by the shooting angle calculation unit of the control device, a shooting direction of the depth camera in the field of view according to the three-dimensional image; A dynamic object recognition unit of the device corrects the three-dimensional image according to the shooting direction to obtain a corrected three-dimensional image, and identifies the dynamic object in the corrected three-dimensional image; and the detection range determining unit defines a three-dimensional image by the control device The platform in the platform, and correcting the platform according to the shooting direction to obtain the detection range; and a determining unit of the control device compares the relationship between the detection range and the dynamic object to determine whether to generate the control signal To trigger the drop buffer. The method for operating a fall prevention device according to claim 16, wherein the step of obtaining the photographing direction of the depth camera in the field of view comprises: finding a plane from the three-dimensional image; and calculating the plane according to the plane The shooting direction of the depth camera. The operator of the fall prevention device as described in claim 16 of the patent application scope The method of modifying the three-dimensional image to obtain the corrected three-dimensional image comprises: performing global coordinate conversion on the three-dimensional image to obtain the corrected three-dimensional image. The method for operating a fall prevention device according to claim 16, wherein the step of identifying the dynamic object in the corrected three-dimensional image comprises: identifying a position of a head, a body, an upper limb or a lower limb of the dynamic object. The method for operating a fall prevention device according to claim 16, wherein the step of obtaining the detection range comprises: determining, according to the platform defined by the user from the three-dimensional image, the platform The level of the detection range; and the world coordinate conversion of the position of the detection range. The method for preventing a fall device according to claim 16, wherein the step of obtaining the detection range comprises: determining a level of a plurality of planes in the three-dimensional image; and determining a level from the planes of the planes; One of the planes defines one of the planes as the platform; and the detection range is obtained by modifying the platform according to the shooting direction. The method for operating a fall prevention device according to claim 21, wherein the step of modifying the platform to obtain the detection range comprises: performing global coordinate conversion on a position of the detection range. The method for operating a fall prevention device according to claim 16, wherein the determining unit compares the detection range with the dynamic object at a world coordinate The correlation is to determine whether the control signal is generated to trigger the drop buffer. The method for preventing a fall device according to claim 23, wherein when the position of the body of the dynamic object is within the detection range, and the position of the center of gravity of the body of the dynamic object exceeds the detection range, and When the height of the head and the body of the dynamic object is lower than a lying range, the control unit generates the control signal to trigger the drop buffer device. The method of operating a fall prevention device according to claim 15, wherein the drop buffer device comprises a self-propelled cushion or an airbag. The operation method of the fall prevention device according to claim 15, further comprising: when the fall warning device of the fall prevention device is triggered, the fall warning device immediately prevents the dynamic object from falling from the platform When the position of the body of the dynamic object is within the detection range, and the head of the dynamic object exceeds the detection range, and the height of the head of the dynamic object is lower than a lying range, generated by the determining unit The control signal triggers the drop warning device. The method of operating a fall prevention device according to claim 26, wherein the fall warning device comprises a jack-up bed guard rail; and wherein the drop warning device includes an alarm for issuing an alarm to indicate the Dynamic objects are about to fall from the platform.